I have always been interested in errors. In particular, I have often wondered about the commonly accepted wisdom that a batter who reaches on an error simply makes an out and deserves no credit for the favorable outcome, that he just happened to be the lucky recipient of a bad play by the opposition. This argument has always reminded me of a much earlier one concerning bases on balls. In the bygone days before Moneyball and Alan Roth, people would argue that batters deserved no credit for walks. They simply had the fortune to be standing at the plate when the pitcher had a lapse of control. Of course, we now know that this argument is ridiculous. Batters arguably have more to do with walks than pitchers, and while I wouldn't make the same statement with respect to reaching base on errors, I do wonder if this is also a skill that some batters possess and others don't.
So the first question I wanted to answer in this article is a simple one: do some batters reach base due to errors much more often than others. On one level the answer to this question seems obvious: since batters strike out, fly out and ground out at different rates, and since each of these three ways of making an out have very different associated error rates, batters who ground out in a high percentage of their at-bats should reach base on errors more often than batters who predominately strike out and fly out. But there still are a host of other potentially interesting issues I want to explore. Are there significant differences even among similar classes of hitters? Are there situational factors that need to be taken into consideration? For example, if there are much lower error rates with no one on, do lead-off batters reach base this way less frequently than clean-up hitters? How big a factor is batter speed? Or whether the batter bats from the right or the left side?
Another major focus of this article will be on park factors. Do some parks have much higher errors rates than others, due either to the influence of official scorers or environmental factors such as rocky infields, poor lighting and unusual wind currents. We've all been at games where a batter reaches base on a hard-hit ground ball that isn't handled cleanly. We look to the scoreboard to see if the play will be ruled a hit or an error. These are judgement calls that have a great impact on the statistical landscape of baseball, in some cases deciding batting championships, RBI crowns and ERA titles, and yet official scorers do their jobs relatively anonymously. No box scores list official scorers, although their decisions can dramatically affect their contents. Except on rare occasions when a late-inning call affects a no-hitter in progress, official scorers fly well below the radar of most fans.
They are so anonymous that this poses special problems to researchers. Several years ago, Retrosheet decided to create game logs which would contain information on every major league game ever played. There was considerable discussion of what to include and the general rule of thumb was: when in doubt, throw it in. When we were done, we had compiled a list of 161 different pieces of information we wanted to include (for a description of these field, please click here). We did not have a field for official scorers, primarily because we felt this would be an almost impossible field to research. Still, I hope that this article stimulates interest in official scorers and perhaps some intrepid person will actually attempt to collect this information at some point.
This article will attempt to explore all of these issues, although not necessarily in the order above. Before I begin, however, I must admit to some fuzzy terminology. When I talk about a batter reaching base on an error, I'm talking about some things that are not classified as errors. I'm also including a batter who strikes out and reaches first because of a wild pitch or passed ball. I'm including a batter who reaches due to a bad fielder's choice that results in no outs being recorded on the play. In short, I'm interested in any play where no outs are made (except for the cases where the batter or runner gets greedy and is thrown out attempting to take an extra base) and the batter is charged with either a hitless at-bat, sacrifice hit or sacrifice fly.
Note that I am not including catcher's interference in this group. I'm not doing this for two reasons. First of all, these plays aren't counted as at-bats and so are not instances where a batter is charged for a bad thing when a good thing really happened. Of course, one could argue that sacrifice hits and flies are also not bad things, but many people see them that way. Certainly, treating sacrifice flies as at-bats for the purposes of calculating on-base percentage is an admission that these are little different than other productive outs. And except for cases when a pitcher is at the plate or we are in the final inning of a one-run game, sacrifice bunts are often more of an unnecessary evil than a smart move. In addition, catcher interference has some official standing, unlike reached on errors, and is already tabulated and included in official yearly statistical reports.
In this article, I will be examining play-by-play data of games from 1960 to 2004. We do not have play-by-play information for ALL of these games, but we come pretty close. For the number of our missing games by league and team, please see the article on the Value Added method.
Let's start by answering this question in the simplest terms, by ignoring hit type, situation factors, parks - in short, everything that might complicate the analysis. If you just look at the number of times a player's outs turn into errors, do some players have much higher error rates than others?
To answer this, I computed how many outs a player was charged with (including sacrifice hits and flies) as well as how many of those resulted in an error (including poor fielder's choices and strike outs where the batter reached first due to a wild pitch or passed ball). For each year, I also generated an expected number of errors, by multiplying the number of outs by the league average of errors per out. I summed all of these, the player's actual and expected errors, for his career and compared them.
What did I find? Well, among players committing at least 2000 outs in their careers from 1960 to 2004, here are the ones who exceeded their expected errors by the greatest percentage:
Name Outs Err GO% FO% SO% ExErr ErrF
Derek Jeter 3859 114 47.0 27.9 25.2 66.9 1.705
Otis Nixon 3831 124 51.3 30.6 18.1 73.4 1.689
Manny Mota 2631 93 55.7 32.7 11.6 59.8 1.556
Rey Sanchez 3627 102 51.9 34.2 13.9 66.7 1.529
Mickey Stanley 3855 127 47.9 37.5 14.6 83.5 1.520
Bob Horner 2781 89 37.1 44.5 18.4 58.7 1.516
Rondell White 3276 91 42.7 32.7 24.5 60.6 1.502
Joe Girardi 3131 89 48.9 31.7 19.4 59.3 1.500
Wil Cordero 3131 85 36.9 38.8 24.3 58.5 1.453
Willie McGee 5471 161 53.6 23.8 22.6 111.8 1.440
Stan Javier 3794 102 45.8 32.1 22.1 71.5 1.427
Greg Gross 2743 86 52.8 38.1 9.1 60.5 1.422
Cesar Tovar 4130 126 45.1 45.0 9.9 89.8 1.403
Jose Vizcaino 3816 102 48.5 33.8 17.7 72.8 1.402
Deivi Cruz 2916 71 46.6 39.8 13.5 50.7 1.400
Chad Curtis 3039 77 39.7 38.0 22.2 55.1 1.397
Miguel Tejada 3122 75 41.4 38.9 19.7 53.8 1.394
Gary Disarcina 2876 72 52.2 37.1 10.6 51.9 1.389
Scott Fletcher 4029 108 48.4 38.2 13.4 77.8 1.388
Roberto Clemente 4526 146 54.0 25.4 20.6 105.7 1.381
Where: Outs - number of outs made
Err - number of times reached on errors
GO% - percentage of outs that were ground balls
FO% - percentage of outs that were fly balls
SO% - percentage of outs that were strikeouts
ExErr - expected number of errors based on league rates
ErrF - error factor (Err / ExErr)
And the lowest:
Name Outs Err GO% FO% SO% ExErr ErrF Darren Daulton 2792 29 30.6 43.4 26.0 56.2 0.516 Mike Lowell 2264 21 28.5 50.7 20.8 40.5 0.519 Jim Gentile 2169 25 32.7 37.0 30.2 48.2 0.519 Mo Vaughn 3957 37 32.5 31.3 36.1 70.8 0.523 Mike Epstein 2180 25 31.1 39.4 29.6 46.6 0.537 Ernie Whitt 2893 33 38.6 44.4 17.0 57.6 0.573 Bobby Murcer 4967 62 34.3 48.7 16.9 107.0 0.579 Bernie Carbo 2036 26 38.7 31.4 29.9 44.6 0.582 Henry Rodriguez 2272 26 27.8 36.9 35.3 44.1 0.589 Jim Dwyer 2101 26 31.5 49.4 19.1 43.6 0.596 Darrin Fletcher 2918 34 38.1 48.3 13.7 55.5 0.613 Greg Walker 2143 26 36.3 39.5 24.3 42.4 0.613 Carlos Delgado 3656 39 30.3 35.7 34.0 63.5 0.614 Franklin Stubbs 2027 25 27.8 41.3 30.9 40.5 0.617 Sid Bream 2353 30 38.3 42.5 19.1 48.3 0.622 Jason Giambi 3408 37 29.7 43.5 26.9 59.1 0.626 Ken Henderson 3440 48 36.0 41.8 22.2 76.3 0.629 Andy Van Slyke 4222 55 35.6 39.2 25.2 86.5 0.636 Jeromy Burnitz 3617 42 29.8 37.2 33.0 65.8 0.638 Boog Powell 5004 70 39.7 35.8 24.5 108.7 0.644
These are lists of very different types of players. For one thing, the players on the upper list hit a lot more ground balls than those on the lower. Here are the averages of the two groups:
GO% FO% SO%
Lots of Errors 47.4 35.0 17.6
Few Errors 33.4 40.9 25.7
Another thing that seems apparent is that the players on the bottom list tend to be a lot slower than the ones on the top. So it does look as if speed has something to do with the ability to coax errors out of a defense.
Still, there are anomalies. For example, Bob Horner would fit in much better with the players who seldom reach on errors. He's a slow, fly ball hitter. But instead of being surrounded by Jim Gentile and Mo Vaughn, he finds himself in the company of Willie McGee and Cesar Tovar. So how much of this can be explained by simple randomness?
There's probably an elegant mathematical way to determine this, but such an approach is beyond my abilities. Instead, I'm going to resort to the last refuge of the mathematically challenged: simulation. I simulated a random distribution of errors and compared these results to what actually happened. This approach is perhaps best shown by example.
The first season we have play-by-play data for Roy McMillan is 1960. He made 315 outs that season. In the National League that year, batters made 31953 outs and reached on error 728 times, for a rate of .022785 per out. So to simulate a random season, I generated 315 random numbers (one for each out he made) between 0 and 1. If a number was less than .022785, I counted it as an error. I totaled all the simulated errors for that season and then did the same thing for all the seasons we have. When I was done, I had a randomly generated number of "errors" in Roy McMillan's career (or at least that portion of his career for which we have play-by-play data). I then took that randomly generated number and used it, along with all the other player results to compute the variance from the expected value. I did this because I wanted to know if the spread of all the actual values was greater or less than the spread when I randomly generated them.
I repeated this experiment 1000 times. In addition to the variance of the randomly generated values, I also calculated the fewest and most errors randomly generated for each player, along with the number of times one of these totals was either greater than his actual value (for players with error factors higher than one) or less (for those with error factors under one). Again, I'm trying to get a feeling for how likely it is that strictly random forces are at play here. This information isn't going to mean much for players with factors close to one, but it should be more interesting for the outliers like the ones shown above.
What did I find out? Not surprisingly, the spread we see in our data is not random. The variance of the 835 players with 2000 or more outs in our database was 201.55; the next highest value in the 1000 random simulations was 86.27. It is also extremely unlikely that the players on the lists above got there by luck. Here are the results of the 1000 simulations on the players with the highest error factors:
Name Outs Err ExErr ErrF Min Max Exc
Derek Jeter 3859 114 66.9 1.705 43 100 0
Otis Nixon 3831 124 73.4 1.689 49 98 0
Manny Mota 2631 93 59.8 1.556 38 90 0
Rey Sanchez 3627 102 66.7 1.529 40 96 0
Mickey Stanley 3855 127 83.5 1.520 57 110 0
Bob Horner 2781 89 58.7 1.516 33 84 0
Rondell White 3276 91 60.6 1.502 35 85 0
Joe Girardi 3131 89 59.3 1.500 36 86 0
Wil Cordero 3131 85 58.5 1.453 36 82 0
Willie McGee 5471 161 111.8 1.440 82 140 0
Stan Javier 3794 102 71.5 1.427 44 99 0
Greg Gross 2743 86 60.5 1.422 36 87 2
Cesar Tovar 4130 126 89.8 1.403 65 123 0
Jose Vizcaino 3816 102 72.8 1.402 47 103 1
Deivi Cruz 2916 71 50.7 1.400 31 75 6
Chad Curtis 3039 77 55.1 1.397 34 80 3
Miguel Tejada 3122 75 53.8 1.394 34 82 5
Gary Disarcina 2876 72 51.9 1.389 29 75 6
Scott Fletcher 4029 108 77.8 1.388 52 107 0
Roberto Clemente 4526 146 105.7 1.381 72 150 1
Where: Min - the minimum number of random errors in 1000 simulations
Max - the maximum number of random errors in 1000 simulations
Exc - the number of times the random errors exceeded the actual ones
None of the top 11 players on the list had their error total met once in 1000 random simulations. And no one on the list had even a 1% chance of being average in this regard.
Now I mentioned earlier that this is not too surprising. After all, most errors are made on ground balls and it's common knowledge that there are ground ball and fly ball hitters. In the rest of the article we will develop more sophisticated ways of determining the number of times a batter might be expected to reach base on errors.
Yes.
Okay, perhaps we should expand on that answer. What follows is a table with information on the three ways of making outs (ground outs, fly outs and strikeouts) in each of the 24 game situations (where outs go from 0 to 2 and the bases go from empty to full).
MenOn ---- GO ---- ---- FO ---- ---- SO ---- FST Out TOT ERR TOT ERR TOT ERR --- 0 39.4 3.96 39.2 0.36 21.4 0.36 --- 1 39.0 3.84 38.0 0.35 23.0 0.38 --- 2 38.7 3.86 37.0 0.38 24.3 0.42 x-- 0 50.1 5.95 33.5 0.34 16.4 0.00 x-- 1 43.0 5.27 37.9 0.36 19.0 0.00 x-- 2 39.0 3.41 39.1 0.40 21.9 0.27 -x- 0 47.8 6.07 31.8 0.42 20.4 0.40 -x- 1 38.3 4.99 37.3 0.43 24.4 0.30 -x- 2 38.5 4.14 36.3 0.40 25.1 0.35 xx- 0 51.4 7.57 30.9 0.34 17.7 0.00 xx- 1 40.9 6.09 38.1 0.36 21.0 0.00 xx- 2 38.7 3.94 37.6 0.37 23.7 0.30 --x 0 36.7 6.14 39.3 0.42 24.0 0.44 --x 1 37.8 8.82 39.0 0.49 23.2 0.44 --x 2 39.2 3.89 35.9 0.42 25.0 0.36 x-x 0 42.5 8.40 39.3 0.56 18.3 0.00 x-x 1 42.7 8.88 38.7 0.46 18.6 0.00 x-x 2 39.6 3.59 37.7 0.46 22.7 0.24 -xx 0 37.6 6.61 38.6 0.62 23.8 0.31 -xx 1 36.8 9.54 38.0 0.47 25.1 0.41 -xx 2 37.7 4.13 34.9 0.45 27.3 0.29 xxx 0 38.3 7.25 40.2 0.49 21.4 0.00 xxx 1 40.0 7.52 38.8 0.43 21.2 0.00 xxx 2 37.8 4.38 37.6 0.46 24.6 0.28
The number on the right (under "TOT") shows how frequent the out is in that situation. So with no one on and no one out, the batter is out 39.4% of the time on a ground out, 39.2% of the time on a fly out, and 21.4% of the time on a strikeout.
The number on the right (under "ERR") shows how frequent an error is for that type of play in that situation. So with bases loaded and no one out, a batter will be safe on an error 7.25% of the time on a ground out, .49% of the time on a fly out and never on a strikeout (since the catcher does not have to cleanly field a third strike with first base occupied and less than two out).
The first thing to notice is that the error rates are VERY different for different types of plays. Not surprisingly, ground outs results in errors around ten times as often as fly outs, and batters reach base least often on a strikeout, but there are situations (no one on) when the fly out is the least likely play to result in an error.
The next thing of interest is that the frequency of plays vary from situation to situation. Strikeouts are at their highest in all situations when there are two outs. Ground outs spike to more than half of all outs when there is either a man on first or a man on first and second with no outs.
Error rates also vary. For ground outs, the error rate goes from a low of 3.41% (man on first and two outs) to a high of 9.54% (men on second and third and one out). Fly outs error rates go from a low of .34% (no outs and a force at second) to a high of .62% (men on second and third and no outs).
But... some of these differences could be explained by our methodology. We chose to lump sacrifice hits in with other outs, which probably explains the spike in ground ball outs in sacrifice situations. And we also chose to treat failed fielder's choices as errors, even when no error was given. This could explain the differences in ground ball error rates when there are unforced men on base.
So let's see how many of these differences are caused by these decisions. Here is the same tables with sacrifice hits and unsuccessful (but not erroneous) fielder's choices removed:
MenOn ---- GO ---- ---- FO ---- ---- SO ---- FST Out TOT ERR TOT ERR TOT ERR --- 0 39.4 3.96 39.2 0.36 21.4 0.36 --- 1 39.0 3.84 38.0 0.35 23.0 0.38 --- 2 38.7 3.86 37.0 0.38 24.3 0.42 x-- 0 42.0 5.33 38.9 0.34 19.0 0.00 x-- 1 41.1 5.09 39.2 0.36 19.7 0.00 x-- 2 39.0 3.36 39.1 0.40 21.9 0.27 -x- 0 41.4 4.17 35.7 0.42 22.9 0.40 -x- 1 38.2 4.59 37.4 0.43 24.4 0.30 -x- 2 38.5 4.11 36.3 0.40 25.1 0.35 xx- 0 42.2 5.65 36.8 0.34 21.1 0.00 xx- 1 39.8 5.78 38.8 0.36 21.4 0.00 xx- 2 38.7 3.78 37.6 0.37 23.7 0.30 --x 0 36.6 4.82 39.4 0.42 24.0 0.44 --x 1 36.6 5.60 39.8 0.49 23.6 0.44 --x 2 39.2 3.87 35.9 0.42 25.0 0.36 x-x 0 41.1 6.48 40.2 0.56 18.7 0.00 x-x 1 41.4 6.50 39.6 0.46 19.0 0.00 x-x 2 39.6 3.56 37.7 0.46 22.7 0.24 -xx 0 37.5 5.07 38.7 0.62 23.8 0.31 -xx 1 36.4 6.24 38.3 0.47 25.3 0.41 -xx 2 37.7 4.10 34.9 0.45 27.3 0.29 xxx 0 38.3 6.77 40.3 0.49 21.5 0.00 xxx 1 39.7 6.73 39.0 0.43 21.3 0.00 xxx 2 37.8 4.23 37.6 0.46 24.6 0.28
The ground outs rates are much flatter with sacrifices out of the picture and the small peaks that remain are probably caused by unsuccessful attempts to move the runners along. Even removing certain fielder's choice did not eliminate all of the variance in ground out error rates. Error rates in some situations are still around double those in others.
Two things are clear from this analysis. First, we should definitely take into account the type of outs a batter makes before declaring that he has a "talent" for reaching on errors. And secondly, it would be a good idea to consider the context of his outs as well, since expected error rates vary quite a bit from situation to situation.
I've always wondered whether or not certain parks were more "error-friendly" than others. In addition, I wondered whether parks favored some types of outs over others. To determine this, I looked at each team's rates of errors, ground outs, fly outs and strikeouts in the 24 game situations in both their home and road parks. Using their road rates, I computed an expected number of errors, ground outs, fly outs and strikeouts in the home park. I next generated the four factors by dividing the actual home totals by the expected values.
What did I find? Well, there is certainly a fair amount of noise in the data, but something is going on here. As I did with the players, I also ran 1000 random simulations. And as before, the spread in the data is not random. The variance of the 1132 teams in our database was 211.55; the next highest value in the random simulations was 91.28.
Here are the teams with the highest error factors, along with the results in our 1000 simulations:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc
1993 COL N DEN01 1.766 1.107 .906 .963 132 75 43 98 0
1991 ATL N ATL01 1.717 .994 1.066 .909 132 77 46 112 0
1994 NY N NYC17 1.701 .995 1.014 .993 71 42 21 63 0
1981 SF N SFO02 1.684 1.044 .955 .994 80 48 28 70 0
1991 BAL A BAL11 1.664 1.022 .956 1.042 93 56 33 81 0
1998 SF N SFO02 1.659 .958 .973 1.112 76 46 27 71 0
2000 BOS A BOS07 1.625 1.034 .953 1.020 87 54 33 78 0
1989 CHI N CHI11 1.606 1.067 .910 1.035 103 64 40 89 0
1969 DET A DET04 1.596 .980 .944 1.140 104 65 38 94 0
1990 LA N LOS03 1.577 .964 1.038 1.003 117 74 53 103 0
1986 DET A DET04 1.570 .935 1.076 .986 98 62 36 89 0
2002 CLE A CLE08 1.562 1.048 .890 1.100 92 59 37 87 0
1998 DET A DET04 1.553 .944 1.059 1.010 94 61 34 88 0
2004 COL N DEN02 1.545 1.009 1.051 .924 71 46 27 71 1
1993 BOS A BOS07 1.544 1.062 .928 1.021 107 69 44 96 0
1966 CHI N CHI11 1.530 1.011 .983 1.008 128 84 56 113 0
1996 CHI N CHI11 1.497 1.010 .981 1.010 101 67 42 93 0
1988 CHI N CHI11 1.492 1.060 .903 1.068 122 82 54 118 0
1994 COL N DEN01 1.472 1.195 .843 .927 79 54 24 81 1
1980 CAL A ANA01 1.470 1.036 .981 .968 96 65 41 95 0
Where: ERR/F - actual errors in home games divided by expected errors (given road rates)
GO/F - actual ground outs in home games divided by expected ground outs
FO/F - actual fly outs in home games divided by expected fly outs
SO/F - actual strikeouts in home games divided by expected strikeouts
AErr - actual errors in home games
EErr - expected errors in home games (given road rates)
Min - the minimum number of random errors in 1000 simulations
Max - the maximum number of random errors in 1000 simulations
Exc - the number of times the random errors exceeded the actual ones
The last column in the table indicates that these teams are not on this list due to chance. Only twice in all of the 1000 random simulations did any one of these 20 teams randomly meet or exceed the actual number of home errors. So I'm reasonably confident that there were factors in the games played in these parks that led to significantly higher than normal error rates. Environmental factors could be to blame, but the obvious cause would seem to be the official scorer. Clearly, many error/hit decision made by the scorers are not clear-cut and I'm sure we've all been to baseball games where we thought a decision of theirs was overly harsh or lenient.
The teams with the lowest error factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1981 STL N STL09 .567 1.023 .947 1.060 41 72 45 110 0 1994 MIN A MIN03 .590 .966 1.012 1.045 43 73 46 108 0 1987 BOS A BOS07 .638 1.027 1.007 .939 50 78 51 109 0 1986 CHI A CHI10 .641 1.029 .982 .982 54 84 51 114 1 1982 PHI N PHI12 .644 .966 .989 1.099 61 95 68 124 0 1994 ATL N ATL01 .651 .929 1.106 .971 39 60 38 86 1 1985 CIN N CIN08 .655 .997 1.049 .924 66 101 73 135 0 1971 KC A KAN05 .657 1.006 1.042 .911 69 105 78 136 0 1993 TOR A TOR02 .660 1.004 .994 1.004 59 89 61 123 0 1975 HOU N HOU02 .661 1.006 .980 1.022 77 116 81 150 0 2002 ANA A ANA01 .661 1.032 .987 .975 57 86 51 124 2 1993 MON N MON02 .664 1.030 1.000 .943 90 136 97 178 0 1994 STL N STL09 .669 1.001 .947 1.093 39 58 34 87 5 1989 MON N MON02 .670 1.017 .932 1.080 68 102 69 134 0 1963 BAL A BAL11 .670 1.002 .923 1.141 59 88 53 115 1 2001 PHI N PHI12 .672 1.001 .991 1.010 53 79 51 118 4 2004 KC A KAN06 .681 1.027 1.009 .944 67 98 64 130 1 1963 LA N LOS03 .682 1.009 1.007 .976 81 119 88 157 0 1998 BAL A BAL12 .690 .992 1.043 .952 49 71 46 99 2 1976 KC A KAN06 .695 1.002 1.093 .796 72 104 73 133 0
I don't do much with the out factors in this article. I included them simply because I thought they might spark some interest among ballpark researchers. Note that these outs will not affect expected error rates (in other words, parks with high ground out rates will not necessarily have high error factors), since the types of outs are already considered when determining these rates.
The parks with the highest ground out factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1994 COL N DEN01 1.472 1.195 .843 .927 79 54 24 81 1 1970 ATL N ATL01 .776 1.147 .836 .984 80 103 73 133 12 1968 LA N LOS03 .938 1.136 .919 .911 85 91 61 124 272 1976 HOU N HOU02 1.017 1.126 .894 .951 99 97 64 129 419 1969 ATL N ATL01 .723 1.123 .895 .893 77 107 80 139 0 1963 CHI N CHI11 1.057 1.121 .973 .840 113 107 78 137 298 1985 KC A KAN06 .891 1.118 .951 .869 83 93 63 127 160 1995 STL N STL09 .771 1.113 .964 .881 65 84 59 115 16 1983 HOU N HOU02 .923 1.111 .883 1.013 88 95 68 125 222 1993 COL N DEN01 1.766 1.107 .906 .963 132 75 43 98 0 2000 COL N DEN02 1.136 1.107 .915 .953 97 85 51 116 112 1992 SF N SFO02 .842 1.106 .983 .865 77 91 62 122 76 1960 LA N LOS01 1.187 1.102 .905 .982 106 89 58 128 44 1999 LA N LOS03 .802 1.102 .964 .914 74 92 67 126 18 2004 CLE A CLE08 1.079 1.100 .884 1.042 75 70 44 96 276 1971 CHI N CHI11 1.023 1.097 .970 .870 91 89 61 129 427 1962 DET A DET04 .989 1.097 .925 .992 104 105 77 138 457 2003 SD N SAN01 1.046 1.097 .887 1.022 69 66 40 93 369 2003 BOS A BOS07 1.090 1.095 1.004 .865 77 71 46 101 226 1994 CLE A CLE08 .936 1.092 1.000 .846 57 61 38 84 358
The parks with the lowest ground out factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1970 SD N SAN01 1.038 .868 1.178 1.013 99 95 65 128 343 1961 PIT N PIT06 .866 .874 1.106 1.119 87 100 72 142 98 1969 SD N SAN01 .940 .880 1.335 .917 85 90 61 123 272 1995 SEA A SEA02 .910 .886 1.018 1.161 65 71 45 100 247 2001 FLA N MIA01 1.173 .887 1.029 1.133 74 63 37 88 108 1967 PHI N PHI11 .960 .889 1.051 1.136 87 91 59 121 360 2003 FLA N MIA01 .721 .891 1.073 1.082 49 68 40 97 7 2004 SEA A SEA03 1.400 .892 1.038 1.120 79 56 35 81 2 1989 NY N NYC17 1.163 .895 1.052 1.079 101 87 63 113 77 1993 OAK A OAK01 1.021 .896 1.027 1.145 67 66 36 97 428 2000 NY A NYC16 .961 .897 1.103 1.015 69 72 44 99 395 1984 CIN N CIN08 .925 .897 1.087 1.021 81 88 57 122 269 1968 CIN N CIN07 1.360 .899 1.108 1.038 119 87 60 123 1 1968 DET A DET04 1.251 .902 1.002 1.170 90 72 50 97 19 1994 SEA A SEA02 1.192 .905 1.031 1.132 51 43 23 64 110 2001 TB A STP01 .998 .906 1.097 1.009 80 80 55 110 538 1961 KC A KAN05 .999 .909 1.090 1.004 112 112 81 144 562 1969 NY N NYC17 1.102 .909 1.049 1.094 88 80 50 107 195 1980 MON N MON02 .878 .909 1.012 1.195 80 91 62 120 126 1989 PHI N PHI12 .836 .910 1.104 1.010 73 87 57 116 60 1971 PHI N PHI12 1.055 .910 1.009 1.165 78 74 49 102 329
The parks with the highest fly out factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1969 SD N SAN01 .940 .880 1.335 .917 85 90 61 123 272 1970 SD N SAN01 1.038 .868 1.178 1.013 99 95 65 128 343 1972 STL N STL09 1.181 .955 1.152 .855 93 79 53 111 68 1995 CHI A CHI12 .818 .928 1.148 .896 62 76 48 103 49 1966 ATL N ATL01 .752 .914 1.139 .923 70 93 63 133 12 1983 CLE A CLE07 1.107 .942 1.129 .870 88 79 56 108 194 1988 ATL N ATL01 1.063 .936 1.127 .916 84 79 47 109 305 1969 CIN N CIN07 1.089 .917 1.127 .958 100 92 66 125 199 1966 PHI N PHI11 .847 .940 1.120 .937 92 109 77 139 51 1964 NY N NYC17 1.000 .926 1.119 .949 102 102 67 142 525 1997 OAK A OAK01 .943 .959 1.119 .900 62 66 43 103 347 1994 CIN N CIN08 1.085 .954 1.114 .926 51 47 29 69 290 1991 CIN N CIN08 .746 .938 1.113 .927 58 78 51 109 9 1978 PIT N PIT07 .903 .959 1.111 .894 88 97 66 124 180 1968 CIN N CIN07 1.360 .899 1.108 1.038 119 87 60 123 1 1995 ATL N ATL01 1.182 .924 1.107 .995 69 58 36 85 98 1994 ATL N ATL01 .651 .929 1.106 .971 39 60 38 86 1 1961 PIT N PIT06 .866 .874 1.106 1.119 87 100 72 142 98 1988 KC A KAN06 .914 1.036 1.105 .790 73 80 52 107 249 1989 PHI N PHI12 .836 .910 1.104 1.010 73 87 57 116 60
The parks with the lowest fly out factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1969 PIT N PIT06 .820 1.089 .794 1.100 94 115 85 149 19 1998 SD N SAN01 1.024 1.059 .828 1.168 80 78 53 107 462 1970 ATL N ATL01 .776 1.147 .836 .984 80 103 73 133 12 1994 COL N DEN01 1.472 1.195 .843 .927 79 54 24 81 1 1987 HOU N HOU02 .781 1.087 .852 1.102 70 90 62 117 19 1981 TEX A ARL01 .909 1.075 .873 1.089 63 69 44 97 247 1961 LA A LOS02 1.168 1.036 .877 1.162 114 98 70 136 58 1999 HOU N HOU02 1.332 1.000 .882 1.156 93 70 45 96 7 1983 HOU N HOU02 .923 1.111 .883 1.013 88 95 68 125 222 2004 CLE A CLE08 1.079 1.100 .884 1.042 75 70 44 96 276 2003 SD N SAN01 1.046 1.097 .887 1.022 69 66 40 93 369 2002 CLE A CLE08 1.562 1.048 .890 1.100 92 59 37 87 0 1985 CHI N CHI11 1.262 1.071 .892 1.052 104 82 54 108 9 1976 HOU N HOU02 1.017 1.126 .894 .951 99 97 64 129 419 1965 STL N STL07 1.296 1.024 .895 1.154 114 88 59 122 3 1969 ATL N ATL01 .723 1.123 .895 .893 77 107 80 139 0 1982 SF N SFO02 .931 1.026 .902 1.150 88 95 66 128 246 1988 CHI N CHI11 1.492 1.060 .903 1.068 122 82 54 118 0 2004 SF N SFO03 1.134 1.072 .904 1.039 82 72 45 102 151 1983 CHI A CHI10 1.226 1.059 .904 1.064 96 78 51 112 26 1996 CAL A ANA01 1.031 1.024 .904 1.125 90 87 57 114 412
The parks with the highest strikeout factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1978 CAL A ANA01 .997 .976 .932 1.229 92 92 62 122 527 1963 CIN N CIN07 1.053 .949 .937 1.203 97 92 57 123 314 1980 MON N MON02 .878 .909 1.012 1.195 80 91 62 120 126 1979 SEA A SEA02 .897 .961 .968 1.187 83 93 64 122 177 1999 PHI N PHI12 .822 .931 .961 1.177 54 66 41 91 62 1978 SF N SFO02 1.122 .981 .942 1.176 86 77 51 114 161 1986 SD N SAN01 1.126 .969 .943 1.173 80 71 46 106 138 1977 TEX A ARL01 1.050 .965 .952 1.173 99 94 64 121 339 1968 DET A DET04 1.251 .902 1.002 1.170 90 72 50 97 19 2000 PHI N PHI12 .980 .943 .946 1.170 69 70 48 103 459 1998 SD N SAN01 1.024 1.059 .828 1.168 80 78 53 107 462 1971 PHI N PHI12 1.055 .910 1.009 1.165 78 74 49 102 329 1978 NY N NYC17 .983 .999 .927 1.164 82 83 55 116 459 1961 LA A LOS02 1.168 1.036 .877 1.162 114 98 70 136 58 2003 PHI N PHI12 .875 .978 .915 1.162 56 64 40 94 175 1995 SEA A SEA02 .910 .886 1.018 1.161 65 71 45 100 247 1987 SD N SAN01 .992 .939 .980 1.160 85 86 58 112 513 1994 SF N SFO02 1.370 .983 .928 1.160 66 48 29 70 11 1999 SF N SFO02 1.104 .934 .970 1.160 83 75 51 107 203 1999 HOU N HOU02 1.332 1.000 .882 1.156 93 70 45 96 7 1972 PHI N PHI12 .710 .990 .927 1.156 53 75 53 103 1
The appearance of the 1978 Angels on this list seems primarily due a weird home/road split for Nolan Ryan that year. He made 21 of his 31 starts at home that season along with 195 of his league-leading 260 strikeouts.
The parks with the lowest strikeout factors:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1978 KC A KAN06 .751 1.034 1.055 .784 80 107 76 145 6 1988 KC A KAN06 .914 1.036 1.105 .790 73 80 52 107 249 2002 KC A KAN06 .748 1.064 1.070 .792 69 92 65 125 6 1976 KC A KAN06 .695 1.002 1.093 .796 72 104 73 133 0 1977 KC A KAN06 .986 1.035 1.064 .808 106 107 78 142 476 1993 KC A KAN06 1.120 1.045 1.084 .813 73 65 42 94 189 1983 KC A KAN06 .920 1.074 1.006 .815 101 110 68 143 228 2003 KC A KAN06 1.077 1.063 1.047 .818 85 79 52 105 266 1995 KC A KAN06 .938 1.069 1.043 .820 60 64 42 93 344 1987 KC A KAN06 .948 1.058 1.068 .825 76 80 52 111 326 1980 KC A KAN06 1.041 1.081 .990 .836 92 88 50 118 357 1979 TOR A TOR01 .816 1.048 1.017 .839 84 103 72 137 30 1982 KC A KAN06 1.066 1.029 1.045 .840 91 85 61 126 253 1963 CHI N CHI11 1.057 1.121 .973 .840 113 107 78 137 298 1961 BOS A BOS07 .831 1.081 1.007 .842 100 120 88 160 26 1996 KC A KAN06 .852 1.077 1.023 .843 71 83 56 112 102 1997 KC A KAN06 .878 1.048 1.062 .846 58 66 40 90 169 1994 CLE A CLE08 .936 1.092 1.000 .846 57 61 38 84 358 1963 SF N SFO02 1.089 1.070 1.018 .850 122 112 80 149 170 1975 KC A KAN06 .898 1.074 .995 .854 101 113 80 147 152 2000 DET A DET05 1.026 1.070 1.028 .854 76 74 49 100 430
This last chart surprised me. I'm not sure why Kansas City has been such a poor park for strikeouts, but it certainly has been.
Since 4 of the seasons with the highest error factors took place at Wrigley Field, I thought it might be worthwhile to look at the complete data for that park.
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1960 CHI N CHI11 .824 1.036 .984 .957 83 101 68 132 39 1961 CHI N CHI11 1.007 1.069 .937 .986 111 110 76 145 482 1962 CHI N CHI11 .953 1.055 .961 .958 107 112 77 147 321 1963 CHI N CHI11 1.057 1.121 .973 .840 113 107 78 137 298 1964 CHI N CHI11 .916 1.013 .999 .973 94 103 70 138 204 1965 CHI N CHI11 .802 .951 1.039 1.041 97 121 89 155 13 1966 CHI N CHI11 1.530 1.011 .983 1.008 128 84 56 113 0 1967 CHI N CHI11 .950 1.016 .996 .975 103 108 82 140 320 1968 CHI N CHI11 .972 1.019 .966 1.020 98 101 68 132 421 1969 CHI N CHI11 .991 1.062 .969 .941 107 108 76 139 486 1970 CHI N CHI11 .951 1.043 .978 .957 85 89 59 120 343 1971 CHI N CHI11 1.023 1.097 .970 .870 91 89 61 129 427 1972 CHI N CHI11 1.079 1.054 .969 .949 94 87 58 117 228 1973 CHI N CHI11 1.248 1.041 .956 .987 98 79 55 102 11 1974 CHI N CHI11 1.268 .998 .984 1.035 137 108 76 148 4 1975 CHI N CHI11 .962 1.019 .993 .971 108 112 81 147 346 1976 CHI N CHI11 1.020 .986 1.046 .944 91 89 64 117 460 1977 CHI N CHI11 1.154 .981 1.000 1.042 106 92 65 127 71 1978 CHI N CHI11 1.349 1.027 .962 1.015 113 84 57 114 1 1979 CHI N CHI11 .948 1.036 .991 .941 88 93 67 123 357 1980 CHI N CHI11 1.248 .977 1.052 .963 106 85 62 113 11 1981 CHI N CHI11 .785 1.046 .968 .970 65 83 57 115 25 1982 CHI N CHI11 1.119 1.030 .980 .975 96 86 60 122 150 1983 CHI N CHI11 1.054 1.027 .935 1.067 89 84 56 113 334 1984 CHI N CHI11 .739 1.037 .943 1.029 75 102 71 132 2 1985 CHI N CHI11 1.262 1.071 .892 1.052 104 82 54 108 9 1986 CHI N CHI11 .929 1.054 .907 1.069 76 82 56 112 293 1987 CHI N CHI11 1.126 1.072 .915 1.016 97 86 54 117 134 1988 CHI N CHI11 1.492 1.060 .903 1.068 122 82 54 118 0 1989 CHI N CHI11 1.606 1.067 .910 1.035 103 64 40 89 0 1990 CHI N CHI11 .919 1.028 .983 .977 98 107 75 141 209 1991 CHI N CHI11 1.421 1.021 .991 .975 106 75 49 102 0 1992 CHI N CHI11 1.020 1.016 .962 1.036 88 86 52 114 419 1993 CHI N CHI11 .994 1.035 .951 1.016 87 88 61 124 525 1994 CHI N CHI11 .983 1.034 .999 .951 65 66 39 92 453 1995 CHI N CHI11 .950 1.025 .950 1.038 75 79 52 106 359 1996 CHI N CHI11 1.497 1.010 .981 1.010 101 67 42 93 0 1997 CHI N CHI11 1.104 .951 1.017 1.056 83 75 47 107 188 1998 CHI N CHI11 1.020 .964 1.043 .991 73 72 49 98 448 1999 CHI N CHI11 1.000 1.023 .984 .989 88 88 60 119 513 2000 CHI N CHI11 .746 1.077 .915 1.013 63 84 60 110 3 2001 CHI N CHI11 .968 .992 .948 1.072 80 83 57 112 402 2002 CHI N CHI11 1.146 .999 .930 1.086 89 78 54 106 103 2003 CHI N CHI11 .910 .954 .954 1.111 78 86 61 123 241 2004 CHI N CHI11 1.134 1.030 1.025 .934 75 66 46 91 161
There doesn't seem to be any real pattern here. I'm not sure whether or not variable environment factors are coming into play (hot summers, windy conditions) or whether their official scorers were replaced often or inconsistent from one year to the next, but we see some weird things here. For example, from 1988 to 1991 the Cubs had three years with very high error factors (and extremely little chance of these factors occurring randomly) and one year in the middle with a low error factor. Similarly, they had low errors factors in 6 of the 7 years from 1964 to 1970, but in the other year (1966), they had an extremely high error factor. Overall, the Cubs had error factors greater than average 25 times and lower factors 20.
A few other samples might be interesting. The table below shows the factors for the Atlanta Braves from 1966 to 1975.
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1966 ATL N ATL01 .752 .914 1.139 .923 70 93 63 133 12 1967 ATL N ATL01 .734 1.008 1.059 .889 74 101 73 130 1 1968 ATL N ATL01 .853 1.028 .989 .962 61 72 49 101 117 1969 ATL N ATL01 .723 1.123 .895 .893 77 107 80 139 0 1970 ATL N ATL01 .776 1.147 .836 .984 80 103 73 133 12 1971 ATL N ATL01 .966 1.012 1.043 .887 87 90 60 121 409 1972 ATL N ATL01 1.395 .998 1.048 .910 93 67 47 90 0 1973 ATL N ATL01 1.190 1.023 1.000 .955 90 76 51 100 49 1974 ATL N ATL01 .992 1.053 .980 .923 108 109 78 144 484 1975 ATL N ATL01 .957 .973 1.021 1.030 104 109 68 146 354
I don't know, but I suspect something happened in 1971 to affect the errors rates in Fulton County Stadium. From 1966 to 1970, fielders were more than 20% less likely to be charged with an error in Atlanta than they were when the same two teams played in another park. I'm not sure if this is something we could uncover at this late date, but I would love to know who were the official scorers in Atlanta during that decade and if anything changed in 1971 to make their decisions less friendly to the fielders there.
Just about every team's table raises similar questions. Here's another one:
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1987 STL N STL09 .827 1.018 .993 .977 76 92 62 121 46 1988 STL N STL09 .936 1.015 .988 .983 89 95 69 128 310 1989 STL N STL09 .880 1.061 .997 .885 69 78 49 103 161 1990 STL N STL09 .977 1.020 1.037 .895 73 75 49 100 462 1991 STL N STL09 .766 1.056 1.010 .877 70 91 60 122 9 1992 STL N STL09 .810 1.047 .964 .977 62 77 44 109 44 1993 STL N STL09 .746 .996 1.040 .934 74 99 71 128 2 1994 STL N STL09 .669 1.001 .947 1.093 39 58 34 87 5 1995 STL N STL09 .771 1.113 .964 .881 65 84 59 115 16 1996 STL N STL09 .784 .937 1.036 1.052 71 91 63 127 13 1997 STL N STL09 1.115 .992 1.049 .961 76 68 45 93 178 1998 STL N STL09 .935 1.007 1.009 .973 67 72 45 100 315 1999 STL N STL09 .994 .970 1.096 .919 86 87 56 122 528 2000 STL N STL09 1.145 .925 1.039 1.049 78 68 43 96 129 2001 STL N STL09 1.001 1.005 .973 1.024 70 70 44 103 493 2002 STL N STL09 1.102 .935 1.045 1.043 74 67 46 94 207 2003 STL N STL09 1.199 .992 1.027 .971 69 58 37 79 79 2004 STL N STL09 .757 .951 1.035 1.035 61 81 52 107 15
What changed around 1997 to make errors more common in Busch Stadium? From 1966 to 1993, that park had lower than average strikeout rates in 17 of the 18 years - what happened around 1994 to make the park more neutral in that regard?
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1962 SF N SFO02 1.169 1.033 .970 .994 112 96 63 128 42 1963 SF N SFO02 1.089 1.070 1.018 .850 122 112 80 149 170 1964 SF N SFO02 1.168 1.022 1.052 .884 122 104 68 134 38 1965 SF N SFO02 1.099 1.001 1.018 .969 113 103 70 136 163 1966 SF N SFO02 1.253 1.051 .993 .920 123 98 66 129 5 1967 SF N SFO02 1.228 1.020 .984 .987 118 96 67 129 17 1968 SF N SFO02 1.057 .999 .961 1.070 111 105 73 144 273 1969 SF N SFO02 1.276 .998 1.005 1.000 121 95 64 125 6 1970 SF N SFO02 1.333 .995 .953 1.091 115 86 57 116 3 1971 SF N SFO02 1.155 1.037 .947 1.027 106 92 59 125 88 1972 SF N SFO02 1.023 .976 1.038 .977 95 93 60 123 443 1973 SF N SFO02 1.250 .979 1.045 .960 106 85 59 122 13 1974 SF N SFO02 1.155 .952 1.070 .976 115 100 71 131 65 1975 SF N SFO02 1.023 .929 1.012 1.144 96 94 66 122 419 1976 SF N SFO02 1.270 .924 1.024 1.149 110 87 60 118 13 1977 SF N SFO02 1.284 .998 1.009 .988 105 82 56 110 6 1978 SF N SFO02 1.122 .981 .942 1.176 86 77 51 114 161 1979 SF N SFO02 1.178 1.044 .914 1.077 123 104 74 142 45 1980 SF N SFO02 1.422 .994 .978 1.059 115 81 54 112 0 1981 SF N SFO02 1.684 1.044 .955 .994 80 48 28 70 0 1982 SF N SFO02 .931 1.026 .902 1.150 88 95 66 128 246 1983 SF N SFO02 1.064 .997 1.008 .991 106 100 75 131 264 1984 SF N SFO02 1.096 .928 1.058 1.042 108 99 66 132 167 1985 SF N SFO02 1.286 .930 1.083 1.002 105 82 50 113 9 1986 SF N SFO02 .873 1.016 1.012 .958 93 106 76 141 91 1987 SF N SFO02 .913 1.034 .999 .946 87 95 70 130 200 1988 SF N SFO02 .964 1.065 .906 1.057 88 91 60 122 384 1989 SF N SFO02 1.131 1.038 .972 .986 88 78 52 112 142 1990 SF N SFO02 .986 1.027 .949 1.042 85 86 59 117 469 1991 SF N SFO02 .942 1.047 .963 .983 88 93 61 122 297 1992 SF N SFO02 .842 1.106 .983 .865 77 91 62 122 76 1993 SF N SFO02 .817 1.039 .960 .999 88 108 75 145 20
For 20 straight years, errors were more common in Candlestick park than elsewhere. What happened around 1986 to dramatically change this trend?
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1973 KC A KAN06 .888 1.009 1.007 .958 87 98 66 131 150 1974 KC A KAN06 1.144 .966 1.078 .911 101 88 59 120 112 1975 KC A KAN06 .898 1.074 .995 .854 101 113 80 147 152 1976 KC A KAN06 .695 1.002 1.093 .796 72 104 73 133 0 1977 KC A KAN06 .986 1.035 1.064 .808 106 107 78 142 476 1978 KC A KAN06 .751 1.034 1.055 .784 80 107 76 145 6 1979 KC A KAN06 .945 .951 1.085 .913 98 104 70 141 302 1980 KC A KAN06 1.041 1.081 .990 .836 92 88 50 118 357 1981 KC A KAN06 1.277 1.008 1.036 .886 54 42 25 63 46 1982 KC A KAN06 1.066 1.029 1.045 .840 91 85 61 126 253 1983 KC A KAN06 .920 1.074 1.006 .815 101 110 68 143 228 1984 KC A KAN06 .963 1.048 .997 .905 81 84 55 115 407 1985 KC A KAN06 .891 1.118 .951 .869 83 93 63 127 160 1986 KC A KAN06 .991 1.022 1.046 .881 78 79 55 109 492 1987 KC A KAN06 .948 1.058 1.068 .825 76 80 52 111 326 1988 KC A KAN06 .914 1.036 1.105 .790 73 80 52 107 249 1989 KC A KAN06 .721 1.029 1.007 .939 57 79 56 106 4 1990 KC A KAN06 1.057 1.028 1.009 .939 91 86 57 115 301 1991 KC A KAN06 .965 1.066 1.006 .889 74 77 48 107 399 1992 KC A KAN06 1.186 1.004 1.048 .898 92 78 52 108 60 1993 KC A KAN06 1.120 1.045 1.084 .813 73 65 42 94 189 1994 KC A KAN06 .957 1.006 1.071 .891 49 51 31 77 435 1995 KC A KAN06 .938 1.069 1.043 .820 60 64 42 93 344 1996 KC A KAN06 .852 1.077 1.023 .843 71 83 56 112 102 1997 KC A KAN06 .878 1.048 1.062 .846 58 66 40 90 169 1998 KC A KAN06 1.250 .969 1.061 .958 86 69 39 98 29 1999 KC A KAN06 .765 .997 1.046 .929 67 88 59 116 16 2000 KC A KAN06 .795 1.042 1.000 .926 74 93 63 123 17 2001 KC A KAN06 1.143 1.027 .989 .967 91 80 50 110 107 2002 KC A KAN06 .748 1.064 1.070 .792 69 92 65 125 6 2003 KC A KAN06 1.077 1.063 1.047 .818 85 79 52 105 266 2004 KC A KAN06 .681 1.027 1.009 .944 67 98 64 130 1
While primarily a fielder-friendly environment (with lower than average error factors in 22 of 32 years), the most striking factor about Kauffman Stadium is its effect on strikeouts. Not once has it been a favorable park for them.
Year Team Park ERR/F GO/F FO/F SO/F AErr EErr Min Max Exc 1993 COL N DEN01 1.766 1.107 .906 .963 132 75 43 98 0 1994 COL N DEN01 1.472 1.195 .843 .927 79 54 24 81 1 1995 COL N DEN02 1.102 1.022 1.010 .946 79 72 45 99 219 1996 COL N DEN02 1.095 1.076 .982 .902 92 84 55 114 206 1997 COL N DEN02 1.324 1.065 1.014 .874 89 67 40 91 3 1998 COL N DEN02 1.020 1.027 1.049 .885 67 66 40 97 453 1999 COL N DEN02 .779 1.020 1.036 .914 70 90 62 120 14 2000 COL N DEN02 1.136 1.107 .915 .953 97 85 51 116 112 2001 COL N DEN02 .757 1.069 .965 .952 63 83 56 112 10 2002 COL N DEN02 .998 1.058 .970 .949 70 70 47 95 514 2003 COL N DEN02 .858 1.045 1.004 .927 75 87 56 117 94 2004 COL N DEN02 1.545 1.009 1.051 .924 71 46 27 71 1
I included the Rockies data because I think it's interesting. I would like to know what happened to decrease error rates around 1999 and what happened to cause them to skyrocket last year. It's also not too surprising to note that Denver has always been a tough place to get a strikeout in.
The complete data is here.
For many of the questions above, the answers will probably be: "I don't know" or "Nothing", but I do think it's clear we need to take the park into account when determining expected error rates. I also think that, given the variation in much of this data, we need to average those rates over a three-year period.
I will only average the data for the prior and subsequent seasons if the team played the majority of its home games in the same park and I will weight the current year twice as heavily as the surrounding ones.
The averaged data is here.
So it looks like we need to adjust the simple approach used at the beginning of the article to take into account the type of outs, situations and parks a batter hits in. I'd like to start by factoring in the type of outs and situations. After doing this, the following players have the highest error factors (again, with a minimum of 2000 outs):
Name B SPD Outs Err GO% FO% SO% ExErr ErrF A/Sit SitF
Bob Horner R 3.2 2781 89 37.1 44.5 18.4 58.7 1.516 52.8 1.687
Gene Tenace R 3.4 3390 82 27.2 43.4 29.4 73.3 1.119 52.1 1.574
Glenn Hubbard R 4.9 3466 100 35.4 46.1 18.5 72.7 1.376 63.6 1.573
Ron Kittle R 3.7 2091 43 24.8 39.6 35.6 41.7 1.031 28.2 1.527
Wil Cordero R 4.9 3131 85 36.9 38.8 24.3 58.5 1.453 55.9 1.521
Robby Thompson R 6.1 3543 96 35.3 36.8 27.9 72.2 1.330 65.5 1.466
Jack Clark R 4.6 5113 129 32.7 39.1 28.2 105.0 1.228 88.1 1.464
Kevin McReynolds R 5.7 4064 102 32.8 49.8 17.4 83.2 1.226 69.7 1.463
Pete Incaviglia R 4.5 3229 71 29.8 30.7 39.5 62.6 1.135 49.2 1.442
Jeff Blauser R 5.4 3424 87 35.1 37.5 27.4 68.2 1.275 60.4 1.440
Jim Wynn R 5.7 4991 134 32.6 39.4 28.0 113.7 1.179 93.8 1.429
Reggie Sanders R 7.1 4116 92 30.9 34.2 34.9 77.6 1.185 64.5 1.426
Derek Jeter R 6.5 3859 114 47.0 27.9 25.2 66.9 1.705 80.1 1.424
Otis Nixon B 7.8 3831 124 51.3 30.6 18.1 73.4 1.689 88.1 1.407
Dick Schofield R 6.2 3445 89 37.2 42.9 19.9 66.3 1.343 63.7 1.397
Rondell White R 5.4 3276 91 42.7 32.7 24.5 60.6 1.502 65.9 1.381
Gerald Williams R 6.1 2308 56 37.3 40.0 22.7 41.1 1.363 40.6 1.379
Johnny Bench R 3.7 5702 150 35.1 42.5 22.4 125.1 1.199 109.7 1.367
Glenn Davis R 4.1 2799 69 35.7 42.4 21.9 56.8 1.214 50.5 1.366
Chad Curtis R 6.1 3039 77 39.7 38.0 22.2 55.1 1.397 56.4 1.366
Where: B - batter handedness ('R' - right, 'L' - left, 'B' - switch-hitter)
SPD - batter speed score
A/Sit - expected number of errors adjusted by out type and situation
SitF - adjusted error factor (Err / A/Sit)
Considering that Bob Horner was the only fly-ball hitter on the earlier list, it not too surprising that he jumps to the top of the class once we take the types of outs into consideration.
Those players with the lowest factors:
Name B SPD Outs Err GO% FO% SO% ExErr ErrF A/Sit SitF Ernie Whitt L 3.1 2893 33 38.6 44.4 17.0 57.6 .573 55.9 .590 Mo Vaughn L 3.2 3957 37 32.5 31.3 36.1 70.8 .523 61.7 .600 Jim Gentile L 2.9 2169 25 32.7 37.0 30.2 48.2 .519 41.0 .610 Bernie Carbo L 4.0 2036 26 38.7 31.4 29.9 44.6 .582 41.4 .628 Darrin Fletcher L 2.0 2918 34 38.1 48.3 13.7 55.5 .613 53.4 .637 Darren Daulton L 5.6 2792 29 30.6 43.4 26.0 56.2 .516 44.7 .649 Sid Bream L 3.7 2353 30 38.3 42.5 19.1 48.3 .622 46.1 .651 Boog Powell L 2.6 5004 70 39.7 35.8 24.5 108.7 .644 107.2 .653 Greg Brock L 4.1 2452 32 40.2 40.7 19.1 49.5 .646 48.6 .658 Greg Walker L 4.3 2143 26 36.3 39.5 24.3 42.4 .613 39.1 .665 Bobby Murcer L 5.2 4967 62 34.3 48.7 16.9 107.0 .579 92.9 .667 Lee Thomas L 4.8 2466 36 38.8 45.4 15.8 54.4 .662 53.9 .668 Mike Lowell R 4.1 2264 21 28.5 50.7 20.8 40.5 .519 31.4 .668 Mike Epstein L 3.5 2180 25 31.1 39.4 29.6 46.6 .537 37.0 .675 Travis Lee L 4.5 2247 28 39.3 35.5 25.1 40.5 .692 41.4 .676 Delino DeShields L 7.5 4330 60 43.8 31.7 24.5 83.4 .719 88.0 .682 Ed Kirkpatrick L 4.5 2707 39 39.6 41.3 19.1 58.8 .663 57.2 .682 Joe Orsulak L 5.9 3200 47 45.1 42.3 12.6 63.2 .744 67.8 .693 Andy Van Slyke L 7.4 4222 55 35.6 39.2 25.2 86.5 .636 78.9 .697 Clay Dalrymple L 3.1 2413 42 42.7 40.6 16.7 57.8 .726 60.3 .697
In addition to the adjusted error number and factor, I also added the following pieces of information to the chart: what side of the plate each batter hit from and the batter speed. The batter speed is derived using Bill James' speed scores. We'll discuss these in more detail shortly.
After adjusting these numbers for the parks each player hit in, we have the following leaders:
Name B SPD Outs Err GO% FO% SO% A/Sit SitF IPF A/Prk PrkF
Bob Horner R 3.2 2781 89 37.1 44.5 18.4 52.8 1.687 1.041 54.9 1.620
Gene Tenace R 3.4 3390 82 27.2 43.4 29.4 52.1 1.574 1.018 53.0 1.546
Wil Cordero R 4.9 3131 85 36.9 38.8 24.3 55.9 1.521 .995 55.6 1.529
Glenn Hubbard R 4.9 3466 100 35.4 46.1 18.5 63.6 1.573 1.041 66.2 1.510
Ron Kittle R 3.7 2091 43 24.8 39.6 35.6 28.2 1.527 1.039 29.3 1.470
Robby Thompson R 6.1 3543 96 35.3 36.8 27.9 65.5 1.466 1.011 66.2 1.450
Pete Incaviglia R 4.5 3229 71 29.8 30.7 39.5 49.2 1.442 .999 49.2 1.444
Glenn Davis R 4.1 2799 69 35.7 42.4 21.9 50.5 1.366 .964 48.7 1.416
Reggie Sanders R 7.1 4116 92 30.9 34.2 34.9 64.5 1.426 1.008 65.0 1.415
Rondell White R 5.4 3276 91 42.7 32.7 24.5 65.9 1.381 .983 64.8 1.405
Jim Wynn R 5.7 4991 134 32.6 39.4 28.0 93.8 1.429 1.020 95.6 1.401
Dick Schofield R 6.2 3445 89 37.2 42.9 19.9 63.7 1.397 .997 63.6 1.400
Derek Jeter R 6.5 3859 114 47.0 27.9 25.2 80.1 1.424 1.018 81.5 1.399
Kevin McReynolds R 5.7 4064 102 32.8 49.8 17.4 69.7 1.463 1.054 73.4 1.389
Jack Clark R 4.6 5113 129 32.7 39.1 28.2 88.1 1.464 1.067 94.0 1.372
Johnny Bench R 3.7 5702 150 35.1 42.5 22.4 109.7 1.367 1.000 109.7 1.367
Nate Colbert R 4.8 2583 63 34.3 31.2 34.5 47.6 1.323 .975 46.5 1.356
Adrian Beltre R 4.7 2556 58 36.4 40.5 23.1 43.8 1.325 .977 42.8 1.356
Roberto Kelly R 6.3 3470 87 38.9 36.2 24.8 64.2 1.355 1.001 64.3 1.354
Garry Maddox R 6.9 4614 123 37.5 45.6 16.9 91.1 1.350 1.004 91.5 1.345
Where: IPF - individual park factor
A/Prk - expected number of errors adjusted by park
PrkF - adjusted error factor (Err / A/Prk)
And the following low-marks:
Name B SPD Outs Err GO% FO% SO% A/Sit SitF IPF A/Prk PrkF Mo Vaughn L 3.2 3957 37 32.5 31.3 36.1 61.7 .600 1.064 65.6 .564 Ernie Whitt L 3.1 2893 33 38.6 44.4 17.0 55.9 .590 .992 55.5 .595 Sid Bream L 3.7 2353 30 38.3 42.5 19.1 46.1 .651 1.060 48.8 .614 Bernie Carbo L 4.0 2036 26 38.7 31.4 29.9 41.4 .628 1.019 42.2 .617 Jim Gentile L 2.9 2169 25 32.7 37.0 30.2 41.0 .610 .975 39.9 .626 Greg Brock L 4.1 2452 32 40.2 40.7 19.1 48.6 .658 1.050 51.1 .627 Jason Varitek B 3.4 2019 24 36.5 35.0 28.5 34.0 .706 1.108 37.7 .637 Boog Powell L 2.6 5004 70 39.7 35.8 24.5 107.2 .653 1.019 109.2 .641 Mike Epstein L 3.5 2180 25 31.1 39.4 29.6 37.0 .675 1.039 38.5 .649 Joe Orsulak L 5.9 3200 47 45.1 42.3 12.6 67.8 .693 1.067 72.3 .650 Darrin Fletcher L 2.0 2918 34 38.1 48.3 13.7 53.4 .637 .972 51.9 .655 Mike Lowell R 4.1 2264 21 28.5 50.7 20.8 31.4 .668 1.018 32.0 .657 Lee Thomas L 4.8 2466 36 38.8 45.4 15.8 53.9 .668 1.016 54.7 .658 Bobby Murcer L 5.2 4967 62 34.3 48.7 16.9 92.9 .667 1.011 93.9 .660 Darren Daulton L 5.6 2792 29 30.6 43.4 26.0 44.7 .649 .982 43.9 .661 Greg Walker L 4.3 2143 26 36.3 39.5 24.3 39.1 .665 1.001 39.1 .665 Travis Lee L 4.5 2247 28 39.3 35.5 25.1 41.4 .676 .997 41.3 .678 Ken Henderson B 5.0 3440 48 36.0 41.8 22.2 67.3 .713 1.048 70.6 .680 Ed Kirkpatrick L 4.5 2707 39 39.6 41.3 19.1 57.2 .682 .988 56.5 .690 Andy Van Slyke L 7.4 4222 55 35.6 39.2 25.2 78.9 .697 1.008 79.6 .691
Clearly, hitting from the right-hand side is a huge factor in reaching on miscues, which makes sense since they hit more frequently to the left side of the infield where there is less margin for error on ground balls. Among the batters with 2000 or more outs, the average adjusted error factors were 1.080 for the 463 right-handed batters, .952 for the 116 switch-hitters, and only .874 for the 256 lefties.
Speed is also a big asset. I took the righties, lefties and switch-hitters and broke each of these groups into ten sections, sorted by their adjusted error factors. The average speed scores for the players in each group:
Type # 1 2 3 4 5 6 7 8 9 10
Right 46 5.39 5.20 4.81 4.87 4.74 4.66 4.86 4.45 4.70 4.49
Switch 11 6.86 6.42 5.71 6.45 5.52 5.95 5.86 5.60 4.55 5.25
Left 25 6.10 5.63 5.58 5.39 5.50 4.95 5.11 4.50 4.26 4.10
Where: # - the size of each group. Any left-over players were added to the last group.
1 - the group with the highest adjusted error factors
2 - the group with the second highest adjusted error factors
and so on
I wanted to make one more adjustment. Since the handedness of the batter makes such a big difference, I wanted to adjust for this in order to see if some players hit balls that were harder to field cleanly, independent of the type of balls they hit (ground outs, fly outs and strikeouts), the situations and the parks they batted in, and whether they hit from the right or left side of the plate.
So I computed these factors for each league and then averaged them over a three-year period. The unaveraged factors are here and the averaged factors are here.
After making this last adjustment, the players with the highest error factors were quite a bit different than before:
Name B SPD Outs Err GO% FO% SO% A/Prk PrkF HndF A/Hnd TErrF
Bob Horner R 3.2 2781 89 37.1 44.5 18.4 54.9 1.620 1.069 58.7 1.516
Gene Tenace R 3.4 3390 82 27.2 43.4 29.4 53.0 1.546 1.045 55.4 1.479
Wil Cordero R 4.9 3131 85 36.9 38.8 24.3 55.6 1.529 1.077 59.8 1.421
Glenn Hubbard R 4.9 3466 100 35.4 46.1 18.5 66.2 1.510 1.066 70.5 1.417
Ron Kittle R 3.7 2091 43 24.8 39.6 35.6 29.3 1.470 1.040 30.4 1.413
Willie Crawford L 5.8 2558 66 39.1 35.1 25.8 52.1 1.267 .911 47.5 1.390
Rusty Greer L 5.3 2713 61 39.7 39.9 20.5 50.3 1.212 .880 44.3 1.377
Pete Incaviglia R 4.5 3229 71 29.8 30.7 39.5 49.2 1.444 1.056 51.9 1.367
Todd Helton L 4.0 2728 59 37.4 42.7 19.9 49.1 1.201 .882 43.3 1.362
Otis Nixon B 7.8 3831 124 51.3 30.6 18.1 92.5 1.340 .989 91.5 1.355
Robby Thompson R 6.1 3543 96 35.3 36.8 27.9 66.2 1.450 1.074 71.1 1.350
Greg Gross L 5.1 2743 86 52.8 38.1 9.1 70.5 1.220 .906 63.8 1.347
Jim Wynn R 5.7 4991 134 32.6 39.4 28.0 95.6 1.401 1.049 100.3 1.336
Reggie Sanders R 7.1 4116 92 30.9 34.2 34.9 65.0 1.415 1.059 68.8 1.336
Glenn Davis R 4.1 2799 69 35.7 42.4 21.9 48.7 1.416 1.063 51.8 1.333
Dick Schofield R 6.2 3445 89 37.2 42.9 19.9 63.6 1.400 1.052 66.8 1.332
Rondell White R 5.4 3276 91 42.7 32.7 24.5 64.8 1.405 1.057 68.5 1.329
Dave Hollins B 4.8 2511 57 35.8 36.8 27.4 46.0 1.240 .948 43.6 1.309
Eddie Bressoud R 5.5 2235 57 32.2 40.5 27.2 42.5 1.342 1.028 43.7 1.305
Kevin McReynolds R 5.7 4064 102 32.8 49.8 17.4 73.4 1.389 1.068 78.4 1.301
Where: HndF - individual handedness factor
A/Hnd - expected number of errors adjusted by handedness
TErrF - total error factor (Err / A/Hnd)
The other end of the list:
Name B SPD Outs Err GO% FO% SO% A/Prk PrkF HndF A/Hnd TErrF Mike Lowell R 4.1 2264 21 28.5 50.7 20.8 32.0 .657 1.052 33.7 .624 Mo Vaughn L 3.2 3957 37 32.5 31.3 36.1 65.6 .564 .873 57.3 .646 Ernie Whitt L 3.1 2893 33 38.6 44.4 17.0 55.5 .595 .906 50.2 .657 Jason Varitek B 3.4 2019 24 36.5 35.0 28.5 37.7 .637 .955 36.0 .667 Jim Gentile L 2.9 2169 25 32.7 37.0 30.2 39.9 .626 .937 37.4 .668 Ken Henderson B 5.0 3440 48 36.0 41.8 22.2 70.6 .680 1.002 70.7 .679 Bernie Carbo L 4.0 2036 26 38.7 31.4 29.9 42.2 .617 .907 38.2 .680 Felix Fermin R 4.5 2164 44 62.6 30.6 6.8 60.8 .723 1.057 64.3 .684 Rick Dempsey R 2.9 3704 53 36.1 44.1 19.9 72.6 .730 1.056 76.7 .691 Preston Wilson R 5.6 2169 28 33.6 29.6 36.8 38.3 .730 1.054 40.4 .693 Larry Herndon R 6.2 3613 58 41.0 37.1 21.9 79.3 .732 1.049 83.2 .697 Boog Powell L 2.6 5004 70 39.7 35.8 24.5 109.2 .641 .912 99.5 .703 Lee Thomas L 4.8 2466 36 38.8 45.4 15.8 54.7 .658 .935 51.2 .703 Greg Brock L 4.1 2452 32 40.2 40.7 19.1 51.1 .627 .891 45.5 .704 Dick Groat R 5.0 3170 74 52.8 36.7 10.5 100.2 .738 1.048 105.0 .705 Tim Foli R 5.3 4750 87 47.6 44.0 8.4 116.7 .745 1.050 122.6 .710 Sid Bream L 3.7 2353 30 38.3 42.5 19.1 48.8 .614 .862 42.1 .713 Jose Hernandez R 5.2 3215 44 35.1 24.7 40.2 58.3 .755 1.053 61.4 .717 Jody Davis R 2.4 2765 40 34.9 39.4 25.8 52.4 .764 1.065 55.8 .717 Bobby Murcer L 5.2 4967 62 34.3 48.7 16.9 93.9 .660 .912 85.7 .724
One interesting thing about Horner is that his final adjusted error factor ended up being the same as the one we started out with. He got a big boost (1.516 to 1.687) for being a fly-ball hitter, then saw his rate drop (1.687 to 1.620) because he played in generally error-friendly parks, and then was dropped back to his original rate (1.620 to 1.516) because he's a right-handed hitter.
Once again, the spread we see in our data is not random, although the spread is far less now that we've accounted for many of the things causing it. The variance of the 835 players with 2000 or more outs in our database is now 119.25; the next highest value in the 1000 random simulations was 83.91. It is unlikely (although not as unlikely as before) that the players on those lists above got there by luck. Here are the results of the 1000 simulations on the players with the highest error factors:
Name Outs Err A/Hnd TErrF Min Max Exc Bob Horner 2781 89 58.7 1.516 37 87 0 Gene Tenace 3390 82 55.4 1.479 33 81 0 Wil Cordero 3131 85 59.8 1.421 36 91 1 Glenn Hubbard 3466 100 70.5 1.417 48 105 1 Ron Kittle 2091 43 30.4 1.413 14 51 18 Willie Crawford 2558 66 47.5 1.390 29 76 6 Rusty Greer 2713 61 44.3 1.377 26 66 5 Pete Incaviglia 3229 71 51.9 1.367 32 74 8 Todd Helton 2728 59 43.3 1.362 25 65 9 Otis Nixon 3831 124 91.5 1.355 62 122 0 Robby Thompson 3543 96 71.1 1.350 48 102 2 Greg Gross 2743 86 63.8 1.347 39 90 4 Jim Wynn 4991 134 100.3 1.336 71 132 0 Reggie Sanders 4116 92 68.8 1.336 43 95 3 Glenn Davis 2799 69 51.8 1.333 31 77 14 Dick Schofield 3445 89 66.8 1.332 42 92 8 Rondell White 3276 91 68.5 1.329 44 97 4 Dave Hollins 2511 57 43.6 1.309 23 65 31 Eddie Bressoud 2235 57 43.7 1.305 24 64 31 Kevin McReynolds 4064 102 78.4 1.301 50 108 5
To refresh what this means, the "Min" column represent the lowest error total in 1000 simulations, assuming that the player had only an average error factor; "Max" represents the highest error total, and "Exc" is the number of times the player's actual total was reached or exceeded in the simulations. So Bob Horner and Gene Tenace never had their real life number reached in the simulations, which is pretty strong evidence that they possess some talent not accounted for in all our adjustments to coax errors out of opposition fielders. Two players, Dave Hollins and Eddie Bressoud, had their actual total reached in 3.1% of the simulations, which means that there some chance they are only average in this regard and owe their appearance here to a 30 to 1 long-shot coming in. Still, I'm pretty confident that the overwhelming majority of the players on this list (as well as several others in the next 20) are not simply lucky.
It does seem, however, that if we are making all of these adjustments to attempt to see if players had different abilities to hit into difficult chances, we might want to remove strikeouts from the picture. We've already looked at strikeout rates and seen how they affect a player's ability to reach on an error, but let's see what happens when we ignore them.
So this time, we are ignoring strikeouts, sacrifice attempts and not treating unsuccessful fielder's choice as errors (since they were handled cleanly). How does this change the leader board?
Name B SPD Outs Err GO% FO% A/Prk PrkF A/Hnd TErrF Gene Tenace R 3.4 2371 81 38.0 62.0 46.2 1.755 48.9 1.655 Bob Horner R 3.2 2268 87 45.4 54.6 50.5 1.723 55.0 1.581 Glenn Hubbard R 4.9 2756 92 42.0 58.0 57.2 1.608 62.3 1.478 Rusty Greer L 5.3 2152 58 49.7 50.3 45.8 1.268 39.6 1.465 Wil Cordero R 4.9 2358 79 48.5 51.5 50.3 1.571 55.0 1.436 Rondell White R 5.4 2467 90 56.5 43.5 58.9 1.528 63.2 1.424 Reggie Sanders R 7.1 2664 86 47.2 52.8 56.3 1.529 60.7 1.417 Otis Nixon B 7.8 3070 112 61.8 38.2 83.0 1.350 79.7 1.406 Jim Wynn R 5.7 3565 123 44.8 55.2 83.6 1.472 89.9 1.368 Alex Rodriguez R 6.2 2815 85 49.9 50.1 56.9 1.494 62.3 1.365 Robby Thompson R 6.1 2457 82 46.9 53.1 54.6 1.501 60.2 1.363 Greg Vaughn R 5.5 3169 90 41.7 58.3 61.7 1.459 67.3 1.337 Jeff Blauser R 5.4 2439 83 47.3 52.7 57.3 1.448 62.7 1.323 Todd Helton L 4.0 2184 52 46.7 53.3 45.7 1.137 39.5 1.315 Johnny Bench R 3.7 4415 139 45.1 54.9 98.5 1.411 106.5 1.305 Jack Clark R 4.6 3662 121 45.4 54.6 85.1 1.422 92.7 1.305 Vada Pinson L 7.0 5358 156 49.2 50.8 133.8 1.166 119.9 1.302 Glenn Davis R 4.1 2180 63 45.6 54.4 44.4 1.420 48.4 1.301 Greg Gross L 5.1 2453 74 57.4 42.6 64.2 1.153 57.0 1.297 Gary Gaetti R 4.1 5173 153 47.1 52.9 110.1 1.389 118.3 1.293
Not a tremendous difference, but I do think this focuses more clearly on what we are trying to look at. Some players dropped off the list because removing strikeouts brought them below the 2000 out minimum for inclusion.
The players with the lowest error rates with these plays removed:
Name B SPD Outs Err GO% FO% A/Prk PrkF A/Hnd TErrF Tim Foli R 5.3 4182 70 50.1 49.9 100.6 .696 108.0 .648 Mo Vaughn L 3.2 2528 32 50.9 49.1 57.8 .554 49.2 .651 Rick Dempsey R 2.9 2905 45 43.8 56.2 63.1 .713 68.0 .662 Larry Herndon R 6.2 2788 52 51.9 48.1 72.2 .720 77.4 .672 Dick Groat R 5.0 2784 65 58.2 41.8 88.5 .735 94.1 .691 Joe Orsulak L 5.9 2754 39 50.8 49.2 66.3 .588 55.9 .697 Ernie Whitt L 3.1 2382 31 46.1 53.9 50.3 .617 44.4 .699 Jody Davis R 2.4 2035 36 46.5 53.5 47.1 .764 51.3 .702 Rich Aurilia R 4.4 2311 34 41.4 58.6 45.1 .754 48.2 .706 Boog Powell L 2.6 3752 63 52.2 47.8 98.0 .643 88.7 .710 Manny Sanguillen R 4.4 3270 68 53.6 46.4 87.6 .776 94.4 .721 Luis Alicea B 6.4 2361 31 44.2 55.8 44.7 .693 42.8 .724 Jose Lind R 5.6 2403 47 55.8 44.2 58.5 .804 64.6 .728 Derrel Thomas R 6.5 2941 57 51.2 48.8 72.1 .791 77.4 .736 Bobby Murcer L 5.2 4109 58 41.1 58.9 86.0 .674 78.1 .742 Dick Green R 4.5 2284 44 51.6 48.4 56.0 .785 59.2 .743 Frank Taveras R 7.3 2556 53 55.4 44.6 66.0 .803 70.9 .747 Alan Ashby B 2.4 2533 46 57.3 42.7 63.3 .726 61.4 .749 Eric Young R 6.7 3783 71 52.2 47.8 88.2 .805 94.6 .750 Ed Kirkpatrick L 4.5 2164 35 48.4 51.6 50.9 .687 46.5 .753
Tim Foli, with a very low strikeout rate, moves to the top of this list and Felix Fermin would have been in 3rd place if he had still met the 2000 out requirement.
We shouldn't let all of these adjustments obscure the fact that right-handed ground ball hitters generally reach base on errors a lot more than lefty fly ball hitters. Despite the final results above, Derek Jeter still reaches base a lot more often than any player on these adjusted lists, and one could argue that the most significant list of players we presented in this article is the first, totally unadjusted, one.
Still, I wanted to go through these contortions to see if we could identify two groups of players: one whose batted balls tended to be difficult to handle and one whose outs posed much less of a challenge. Much like the differences in ball park error rates presented above, I don't know if Gene Tenace, Bob Horner and Glenn Hubbard really hit scorching ground-balls or whether Mo Vaughn didn't. Perhaps people who have watched the players on these two lists play more than I have can comment on this. I do know that these differences are unlikely to occur by chance. Even after taking into consideration a host of things that might account for these differences (with the notable exception of batter speed), there still seems to be some significant differences in how difficult each batter is to retire on his outs.
The yearly player data is here.
The career player data is here.
I realize that the title of this article only mentions batters, but I figured it would be an oversight to conclude this piece without a discussion of which pitchers gave up more than their share of errors. This is probably more interesting to current researchers than what I've been talking about so far, in light of recent work (most notably by Voros McCracken and Tom Tippett) on the subject of how much influence pitchers have over the successful disposition of balls in play.
Before getting too far into this, it should be obvious that one big thing pitchers can do to minimize errors is to strikeout as many hitters as they can. Errors rates on strikeouts are extremely low as are errors on fly balls. So we should see a wide disparity between error rates behind different types of pitchers and, at least before any adjustments are made, we do.
The pitchers with the highest error factors:
Name Outs Err GO% FO% SO% ExErr ErrF
Hal Woodeshick 2077 86 56.0 23.4 20.5 50.1 1.715
Bob Locker 2568 91 55.6 21.9 22.5 55.8 1.631
Roger McDowell 3034 99 59.9 22.8 17.3 61.2 1.618
Frank Linzy 2371 83 60.4 24.5 15.1 53.1 1.562
Kent Tekulve 4188 138 54.7 26.7 18.6 89.3 1.546
Rick Honeycutt 6251 195 52.5 30.9 16.6 128.0 1.523
Atlee Hammaker 3157 100 46.5 34.0 19.5 66.0 1.514
Rick Camp 2720 88 53.5 31.5 15.0 58.2 1.512
Mike Fetters 2059 56 46.6 28.2 25.2 37.3 1.500
Randy Jones 5616 185 57.6 29.4 13.1 123.7 1.495
Ted Abernathy 2619 85 51.8 24.9 23.3 57.6 1.475
Andy Hassler 3219 101 51.0 29.5 19.6 69.4 1.456
Jack Aker 2153 68 52.4 28.8 18.8 47.0 1.446
Kip Wells 2324 58 43.7 30.6 25.6 40.5 1.433
Mike Hampton 5762 153 50.1 28.3 21.6 107.0 1.430
Steve Trout 4272 127 55.7 29.0 15.4 89.2 1.424
S. Schoeneweis 2038 50 47.8 32.8 19.3 35.2 1.422
Matt Clement 3376 86 41.9 27.7 30.5 60.7 1.417
Jason Grimsley 2569 65 50.0 26.6 23.4 46.0 1.412
Al Jackson 4002 136 51.0 30.8 18.1 96.9 1.403
Where: Outs - number of outs made
Err - number of times reached on errors
GO% - percentage of outs that were ground balls
FO% - percentage of outs that were fly balls
SO% - percentage of outs that were strikeouts
ExErr - expected number of errors based on league rates
ErrF - error factor (Err / ExErr)
And the lowest:
Name Outs Err GO% FO% SO% ExErr ErrF Eddie Guardado 2165 17 23.9 46.1 30.0 38.1 .447 Jaret Wright 2178 20 38.0 35.9 26.1 38.0 .526 Eric Milton 3495 33 25.4 49.6 25.1 60.9 .542 Jeff Brantley 2497 28 30.0 40.9 29.2 49.8 .562 Robert Person 2610 28 25.7 44.7 29.6 48.2 .581 Jeff Reardon 3335 40 24.6 49.1 26.3 68.2 .587 Dick Radatz 2039 26 24.1 39.4 36.5 44.2 .589 Don Gullett 4016 53 32.4 44.9 22.8 89.6 .591 Bob Buhl 3976 58 43.2 40.3 16.4 97.2 .597 Pat Jarvis 3521 47 38.3 41.3 20.4 77.9 .603 O. Hernandez 2561 27 29.8 42.8 27.5 44.7 .604 Dennis Eckersley 9644 121 30.0 45.1 24.9 199.6 .606 Luis Tiant 10137 134 29.9 46.2 23.8 220.7 .607 Sid Fernandez 5491 69 20.4 47.9 31.7 112.0 .616 Denny McLain 5499 74 31.6 45.0 23.3 119.1 .621 Randy Wolf 3017 34 30.9 40.4 28.7 54.2 .627 Art Mahaffey 2891 45 31.4 46.5 22.1 71.5 .630 Mark Gardner 5094 62 32.9 42.4 24.7 98.1 .632 Gary Nolan 4905 69 33.3 45.5 21.2 108.9 .634 Jim Palmer 11416 162 35.4 45.3 19.4 244.9 .661
As you might expect, the top list is dominated by ground ball pitchers, while the bottom list is filled with the reverse, those who primarily get their outs in the air or by strikeouts.
Adjusting for the type and situation, mixes things up quite a bit.
Name P Outs Err GO% FO% SO% ExErr ErrF A/Sit SitF
Al Hrabosky L 2111 55 28.1 45.9 26.0 46.7 1.177 33.6 1.637
Scott Sanders R 2012 48 32.4 36.2 31.4 38.9 1.233 33.1 1.449
Tim Wakefield R 6008 138 34.1 41.9 24.0 106.4 1.297 95.9 1.439
Atlee Hammaker L 3157 100 46.5 34.0 19.5 66.0 1.514 72.0 1.389
Balor Moore L 2023 55 38.0 38.1 23.9 43.5 1.265 40.1 1.372
Eddie Solomon R 2086 61 43.0 40.8 16.2 44.3 1.378 44.6 1.369
Mark Leiter R 3450 80 34.5 39.6 25.9 65.8 1.216 59.2 1.351
Matt Clement R 3376 86 41.9 27.7 30.5 60.7 1.417 64.4 1.336
Ron Bryant L 2695 76 39.5 41.6 18.9 59.5 1.278 57.1 1.332
Mike Sirotka L 2039 50 41.0 37.7 21.3 36.3 1.378 37.9 1.319
Jason Johnson R 2908 64 37.1 40.7 22.1 49.9 1.282 48.8 1.312
Jim Merritt L 4353 107 34.3 44.3 21.4 96.3 1.111 81.9 1.306
Jeff Fassero L 5663 143 41.5 30.7 27.8 106.2 1.346 110.0 1.300
Ed Halicki R 3151 78 36.3 41.3 22.4 69.1 1.128 60.1 1.297
Tim Worrell R 2664 59 35.1 38.4 26.5 50.2 1.176 45.7 1.290
Ken Kravec L 2512 61 37.1 40.7 22.2 53.9 1.133 47.3 1.290
Kip Wells R 2324 58 43.7 30.6 25.6 40.5 1.433 45.0 1.289
Bill Krueger L 3435 88 40.6 40.8 18.6 66.3 1.327 68.3 1.288
Don Mossi L 2355 57 32.3 47.8 20.0 53.0 1.076 44.3 1.287
Randy Johnson L 9793 178 28.4 29.1 42.5 178.8 .996 138.6 1.284
Where: P - pitcher handedness ('R' - right, 'L' - left)
A/Sit - expected number of errors adjusted by out type and situation
SitF - adjusted error factor (Err / A/Sit)
Only three pitchers on this list were on the earlier one. Randy Johnson actually allowed fewer errors than expected until we took into account his high strikeout rate.
The bottom:
Name P Outs Err GO% FO% SO% ExErr ErrF A/Sit SitF Jaret Wright R 2178 20 38.0 35.9 26.1 38.0 .526 38.1 .525 Bob Buhl R 3976 58 43.2 40.3 16.4 97.2 .597 99.1 .585 Eddie Guardado L 2165 17 23.9 46.1 30.0 38.1 .447 26.9 .633 Sidney Ponson R 3754 45 42.1 36.5 21.4 64.9 .694 69.5 .647 Pat Jarvis R 3521 47 38.3 41.3 20.4 77.9 .603 71.9 .654 Hal Brown R 2011 32 40.4 45.9 13.7 48.3 .663 47.5 .674 Rolando Arrojo R 2003 27 42.7 31.7 25.6 36.0 .750 39.6 .681 John Butcher R 2398 38 48.0 36.9 15.1 48.8 .779 55.6 .683 Ron Taylor R 2313 36 39.6 40.3 20.0 54.2 .664 52.1 .691 Milt Pappas R 8116 123 41.3 40.1 18.6 178.4 .689 177.5 .693 Jeff Brantley R 2497 28 30.0 40.9 29.2 49.8 .562 40.1 .698 S. Kamieniecki R 2770 39 44.7 35.7 19.6 50.0 .781 55.8 .699 Dick Selma R 2418 36 39.9 32.0 28.2 54.0 .667 51.2 .703 Dave McNally L 7841 120 41.3 39.4 19.3 169.6 .707 169.1 .710 Rollie Sheldon R 2099 34 40.9 41.4 17.7 46.9 .725 47.7 .713 Pete Smith R 2966 41 39.1 39.3 21.6 59.3 .692 57.3 .715 Mark Gubicza R 6403 101 47.6 31.0 21.4 122.0 .828 140.3 .720 Brad Radke R 6047 75 38.4 40.6 21.0 105.2 .713 104.0 .721 Mike Mussina R 8199 98 35.5 36.9 27.5 144.6 .678 134.9 .726 Rick Waits L 4059 72 49.7 34.0 16.2 85.6 .841 99.0 .727
Another wholesale turnover, with only five repeats from before. Once again, handedness plays a role here, as the top list has far more than its share of lefties, while the bottom list is dominated by right-handed pitchers.
The park adjustment shouldn't be as dramatic and it isn't.
Name B Outs Err GO% FO% SO% A/Sit SitF IPF A/Prk PrkF Al Hrabosky L 2111 55 28.1 45.9 26.0 33.6 1.637 1.001 33.6 1.636 Balor Moore L 2023 55 38.0 38.1 23.9 40.1 1.372 .957 38.4 1.434 Scott Sanders R 2012 48 32.4 36.2 31.4 33.1 1.449 1.018 33.7 1.422 Mike Sirotka L 2039 50 41.0 37.7 21.3 37.9 1.319 .938 35.6 1.406 Atlee Hammaker L 3157 100 46.5 34.0 19.5 72.0 1.389 1.022 73.6 1.359 Ramon Ortiz R 2592 55 38.0 39.3 22.8 43.8 1.256 .927 40.6 1.354 Mark Leiter R 3450 80 34.5 39.6 25.9 59.2 1.351 1.014 60.0 1.332 Jason Johnson R 2908 64 37.1 40.7 22.1 48.8 1.312 .990 48.3 1.324 Tim Wakefield R 6008 138 34.1 41.9 24.0 95.9 1.439 1.089 104.5 1.321 Jeff Fassero L 5663 143 41.5 30.7 27.8 110.0 1.300 .989 108.8 1.314 Eddie Solomon R 2086 61 43.0 40.8 16.2 44.6 1.369 1.047 46.7 1.307 Matt Clement R 3376 86 41.9 27.7 30.5 64.4 1.336 1.027 66.2 1.300 Don Mossi L 2355 57 32.3 47.8 20.0 44.3 1.287 .996 44.1 1.292 Mark Guthrie L 2818 64 37.9 34.5 27.6 51.0 1.255 .977 49.8 1.285 Randy Johnson L 9793 178 28.4 29.1 42.5 138.6 1.284 1.004 139.2 1.279 Rick Honeycutt L 6251 195 52.5 30.9 16.6 155.5 1.254 .987 153.4 1.271 Bill Krueger L 3435 88 40.6 40.8 18.6 68.3 1.288 1.015 69.3 1.269 Jerry Johnson R 2249 67 43.8 34.4 21.7 52.6 1.275 1.011 53.2 1.261 Ron Robinson R 2311 54 37.1 42.4 20.5 44.3 1.218 .969 43.0 1.257 Paul Byrd R 2649 54 35.4 43.8 20.8 43.5 1.243 .988 43.0 1.257
Name B Outs Err GO% FO% SO% A/Sit SitF IPF A/Prk PrkF Jaret Wright R 2178 20 38.0 35.9 26.1 38.1 .525 1.019 38.8 .515 Bob Buhl R 3976 58 43.2 40.3 16.4 99.1 .585 1.006 99.8 .581 Rolando Arrojo R 2003 27 42.7 31.7 25.6 39.6 .681 1.075 42.6 .634 Eddie Guardado L 2165 17 23.9 46.1 30.0 26.9 .633 .990 26.6 .639 Ron Taylor R 2313 36 39.6 40.3 20.0 52.1 .691 1.071 55.8 .646 Hal Brown R 2011 32 40.4 45.9 13.7 47.5 .674 1.043 49.5 .647 Sidney Ponson R 3754 45 42.1 36.5 21.4 69.5 .647 .995 69.2 .650 Pete Smith R 2966 41 39.1 39.3 21.6 57.3 .715 1.094 62.7 .654 Armando Reynoso R 3077 46 41.9 40.1 18.0 62.2 .740 1.121 69.7 .660 John Butcher R 2398 38 48.0 36.9 15.1 55.6 .683 1.010 56.2 .676 Jamey Wright R 3175 55 50.6 30.4 18.9 74.3 .740 1.078 80.1 .686 Dave McNally L 7841 120 41.3 39.4 19.3 169.1 .710 1.032 174.5 .688 Milt Pappas R 8116 123 41.3 40.1 18.6 177.5 .693 1.004 178.2 .690 Bob Milacki R 2299 33 41.9 41.2 16.8 44.6 .741 1.067 47.5 .694 Mike Mussina R 8199 98 35.5 36.9 27.5 134.9 .726 1.044 140.8 .696 Pat Jarvis R 3521 47 38.3 41.3 20.4 71.9 .654 .933 67.1 .700 Dick Selma R 2418 36 39.9 32.0 28.2 51.2 .703 1.000 51.2 .703 Masato Yoshii R 2166 30 38.8 40.6 20.6 39.8 .754 1.072 42.6 .703 Mark Gardner R 5094 62 32.9 42.4 24.7 84.5 .734 1.039 87.8 .706 S. Kamieniecki R 2770 39 44.7 35.7 19.6 55.8 .699 .990 55.2 .706
Finally, we will adjust this table for the handedness of the pitcher. As before, I computed these factors for each league and then averaged them over a three-year period. The unaveraged factors are here and the averaged factors are here.
This time, the adjustment is not as extreme.
Name B Outs Err GO% FO% SO% A/Prk PrkF HndF A/Hnd TErrF
Al Hrabosky L 2111 55 28.1 45.9 26.0 33.6 1.636 1.036 34.8 1.579
Scott Sanders R 2012 48 32.4 36.2 31.4 33.7 1.422 .992 33.5 1.434
Ramon Ortiz R 2592 55 38.0 39.3 22.8 40.6 1.354 .972 39.5 1.393
Balor Moore L 2023 55 38.0 38.1 23.9 38.4 1.434 1.031 39.6 1.390
Jason Johnson R 2908 64 37.1 40.7 22.1 48.3 1.324 .969 46.8 1.367
Tim Wakefield R 6008 138 34.1 41.9 24.0 104.5 1.321 .972 101.5 1.360
Mark Leiter R 3450 80 34.5 39.6 25.9 60.0 1.332 .980 58.9 1.359
Eddie Solomon R 2086 61 43.0 40.8 16.2 46.7 1.307 .977 45.6 1.338
Matt Clement R 3376 86 41.9 27.7 30.5 66.2 1.300 .992 65.6 1.311
Atlee Hammaker L 3157 100 46.5 34.0 19.5 73.6 1.359 1.047 77.1 1.298
Ron Robinson R 2311 54 37.1 42.4 20.5 43.0 1.257 .974 41.8 1.291
Jerry Johnson R 2249 67 43.8 34.4 21.7 53.2 1.261 .979 52.0 1.288
Bruce Dal Canton R 2707 78 44.5 37.6 17.9 62.4 1.250 .976 60.9 1.281
Dave Morehead R 2399 59 36.8 37.1 26.1 47.4 1.244 .972 46.1 1.279
Mike Fetters R 2059 56 46.6 28.2 25.2 44.8 1.251 .980 43.9 1.276
Paul Byrd R 2649 54 35.4 43.8 20.8 43.0 1.257 .987 42.4 1.274
Eric Plunk R 3341 63 29.0 38.7 32.4 50.5 1.248 .983 49.6 1.269
Tim Worrell R 2664 59 35.1 38.4 26.5 47.5 1.243 .981 46.5 1.268
Larry Sherry R 2044 57 38.7 35.2 26.0 46.0 1.239 .979 45.0 1.266
Reggie Cleveland R 5300 145 41.7 40.8 17.5 117.1 1.239 .986 115.5 1.256
Where: HndF - individual handedness factor
A/Hnd - expected number of errors adjusted by handedness
TErrF - total error factor (Err / A/Hnd)
Name B Outs Err GO% FO% SO% A/Prk PrkF HndF A/Hnd TErrF Jaret Wright R 2178 20 38.0 35.9 26.1 38.8 .515 .974 37.8 .529 Bob Buhl R 3976 58 43.2 40.3 16.4 99.8 .581 .990 98.7 .587 Eddie Guardado L 2165 17 23.9 46.1 30.0 26.6 .639 1.081 28.8 .591 Dave McNally L 7841 120 41.3 39.4 19.3 174.5 .688 1.062 185.4 .647 Ron Taylor R 2313 36 39.6 40.3 20.0 55.8 .646 .987 55.0 .654 Rolando Arrojo R 2003 27 42.7 31.7 25.6 42.6 .634 .969 41.3 .654 Hal Brown R 2011 32 40.4 45.9 13.7 49.5 .647 .973 48.2 .664 Armando Reynoso R 3077 46 41.9 40.1 18.0 69.7 .660 .989 68.9 .668 Pete Smith R 2966 41 39.1 39.3 21.6 62.7 .654 .972 60.9 .673 Sidney Ponson R 3754 45 42.1 36.5 21.4 69.2 .650 .964 66.7 .674 John Butcher R 2398 38 48.0 36.9 15.1 56.2 .676 .975 54.8 .694 Jamey Wright R 3175 55 50.6 30.4 18.9 80.1 .686 .988 79.2 .695 C.C. Sabathia L 2218 27 33.4 39.6 27.1 35.9 .752 1.068 38.3 .704 Milt Pappas R 8116 123 41.3 40.1 18.6 178.2 .690 .979 174.4 .705 Eric Milton L 3495 33 25.4 49.6 25.1 42.7 .773 1.091 46.6 .708 Masato Yoshii R 2166 30 38.8 40.6 20.6 42.6 .703 .992 42.3 .709 Bob Milacki R 2299 33 41.9 41.2 16.8 47.5 .694 .980 46.6 .709 Pat Jarvis R 3521 47 38.3 41.3 20.4 67.1 .700 .984 66.0 .712 Dick Selma R 2418 36 39.9 32.0 28.2 51.2 .703 .982 50.3 .715 Mike Mussina R 8199 98 35.5 36.9 27.5 140.8 .696 .974 137.1 .715
Now two things concern me about this methodology when used with pitchers instead of hitters. First, while a batter puts balls in play against a variety of defenses during the course of a season, a pitcher is stuck (or blessed) with much the same defense in every game. The other important thing to remember is that the pitcher himself is also part of his defense and could be a significant factor in both errors on sacrifice attempts as well as the incidence of strikeout victims reaching base.
So it's not clear to me whether Dave McNally's ability to minimize errors is really a skill we should be attributing to him or to Mark Belanger, the shortstop for many of his starts. Pitchers do move from team to team now and then, and team defenses also change, sometimes dramatically, over time, but these concerns are still there and, at least to me, muddy the water here in a way they didn't for the batters.
The yearly pitcher data is here.
The career pitcher data is here.
It shouldn't be a surprise to anyone that this article raises many questions and comes up with relatively few answers. It does provides some data to back up what most of us already knew: grounders produce more errors than fly outs, righties reach on errors more often than lefties, the speed of a batter affects error rates, and so on. But I feel that the questions it raises are far more interesting than these "answers", and I hope that this article stimulates interest in this somewhat obscure topic and encourages people to investigate some of these open questions. What caused error rates to suddenly drop or rise in certain parks? What caused the fluctuations in some parks' ground out, fly out or strikeout factors? Why were Bob Horner's out so much harder to field cleanly than Mo Vaughn's? Hopefully, this article is a first small step toward answering some of these kinds of questions.