BP Comment Quick Links


March 11, 2007 Prospectus Q&APete Palmer
Credited with introducing OPS as a statistic over 20 years ago, Pete Palmer is the coauthor of The Hidden Game of Baseball and the coeditor of the ESPN Baseball Encyclopedia. David Laurila sat down with Palmer for Baseball Prospectus to talk about the best players not in the Hall of Fame and about the importance of OPS.
 David Laurila: The Veterans Committee didn't elect any new Hall of Famers this year. What are your thoughts on that? Pete Palmer: The setup of the new Veterans Committee makes it very hard to elect anyone. There are many finalists (25 or so) and none stand out far above the others. The writers have it easier because there are usually no more than two or three good candidates each year, plus maybe one or two holdovers from previous years. The Hall of Fame is talking about revising the procedure for the vets now that three elections are in with no one chosen. DL: How do you rate the players who were under consideration? PP: My ratings are all based on our values from the ESPN Encyclopedia, which credits players in wins above the average player for their career, taking into account batting, pitching, fielding, base running, and fielding position. There were 200 initial candidates for the vets, from anyone with 10 or more year's service who had been retired for at least 22 years. Here is a list of my ratings and their status for those eligible:
1. Bill Dahlen (48.5 wins) in initial group, not a finalist My rule of thumb is anyone with 30 or more wins above average is a good candidate, 2030 wins makes a marginal candidate, while below 20 should not be considered. However, the actual Hall is quite different, as you can see from the above list; Bob Johnson didn't even make the top 200 for the vets this time. DL: Among players not currently eligible, Bobby Grich probably stands out as the most deserving of Hall of Fame consideration. Who are the best of the rest? PP: A player gets dropped from the writers' vote if he gets less than five percent of the vote. In the past, anyone with less than 100 votes in the writers' election would not be eligible to be voted on by the vets, but when they reorganized the vet committee for 2003 this rule was rescinded. There were several among the top 200 for the vets who had been lopped off the writers' ballot, although none made it to the finals. Jimmy Wynn did not get a single vote from the writers, (and) Reggie Smith and Norm Cash were under five percent their first year. Here is a list of those dropped by the five percent rule who are not yet eligible for the vet committee:
1. Bobby Grich (51.5) dropped in his first year, eligible in 2009 With the new setup, they may get a second chance, although the rules could change. DL: Alan Schwarz recently wrote about the relative value of OBP and SLG in assessing the meaningfulness of OPS as a statistic. What are your thoughts on that? PP: My method uses them weighted equally. This was first presented in 1984 in The Hidden Game of Baseball, although I had developed the formula maybe 10 years earlier. Adding onbase plus slugging is a shorthand approximation of normalized onbase plus slugging, which in itself is an approximation of linear weights. That is because the values of a walk, single, double, triple and home run using OPS are approximately 1, 2, 3, 4 and 5, which is about the same ratio as in linear weights. A walk adds 1 to onbase average, a single adds 1 to both, a double 1 to onbase average and 2 to slugging, etc. Linear weight has .25 for a walk, .47 for a single, .83, 1.02 and 1.40 for the others. To get the normalized version, you take onbase average divided by the league average plus slugging average divided by the league average all minus 1. Thus a player who was 10 percent higher in each would get a rating of 1.1 plus 1.1 minus 1 or 1.2. This is particularly meaningful because the number of runs produced by this player would be 20 percent better than average, a direct correlation. Doing it this way effectively makes OBP about 20 percent more valuable than slugging because 33 points in OBP are equal to about 40 points in slugging, due to dividing by the league average. I should mention that the league average does not include pitcher hitting, so there is no advantage for National League batters. This is further adjusted by dividing by the park factor, so a 1.2 player with a 1.1 park factor would have a 1.1 rating. Player park factor is approximately park factor plus 1 all over two. Using the simplified OPS skips both these adjustments, but is much easier to calculate. Using simple OPS does not work quite as well, since adding 10 percent in onbase and 10 percent in slugging increases OPS by only 10 percent overall, while run production goes up 20 percent. Branch Rickey, assisted by Alan Roth, came up with a similar formula in 1954, which was onbase average plus threequarters times isolated power (total bases minus hits all over at bats). Earnshaw Cook's DX from the '60s was basically onbase times slugging, while Bill James' runs created was very close to onbase times slugging times atbats. Dick Cramer's batter run average from the 1970s was exactly onbase average times slugging (OXS). He developed it from simulations where all nine men in the lineup had the stats of the player being tested. For teams, OBP times slugging works slightly better, but the difference is small. A team with 15 percent better OBP and slugging would have an OXS rating of 1.15 times 1.15 or 1.32, while OPS would give 1.15 plus 1.15 minus 1 or 1.30. The difference is less than the correlation error, which is around 3 percent. There has been only one team since 1960 outside the .70 to 1.30 normalized OPS range, the 1976 Reds, who were 1.33. For players, though, the OPS formula works better. That is because adding a good player to an average team does not result in as many extra runs as would occur if all nine players were the same. For example, take Babe Ruth with 1.5 OBA and 2.0 slugging. That equates to 3.0 for OXS but only 2.5 for OPS. Assuming 75 runs is average, a team of Ruths might show 225 runs per player, or plus 150, making maybe 2,000 runs for the team in a season, but adding Babe Ruth to an average line up would produce only about 110 runs more than average. Recent studies have shown that a multiplier of somewhere between one and two times OBP plus slugging correlates slightly better with team batting performance, but the difference is negligible. For all teams, 1960 to date, using simple normalized OPS, you get a standard deviation when calculating expected team runs of 23.8. Using the optimal ratio of 1.05 times normalized OBA plus .95 times normalized slugging, the prediction error is still 23.8. Going down to .90/1.10 (a ratio of .8) or up to 1.15/.85 (a ratio of 1.35) gives an error of 24.0, so the results are flat over a wide range. Since the normalized version already has a 1.20 weighting for OBA built in, this suggests that the best value for comparing the two is around 1.30 (1.20 times 1.05/.95). In doing this study, I used normalized team runs per batting out, rather than total runs. Team batting outs can vary by up to 100 or so per year, mainly due to the fact the home team does not bat in the ninth at home if already ahead. DL: John McGraw had an OBP of .466 in 3,924 lifetime atbats. Is he a Hall of Famer based on his playing career alone? With longer careers, what about Ferris Fain and Charlie Keller? PP: McGraw was great at getting on base; in fact he has the highest average in the history of the National League. However, he was a onedimensional player, having no power and not being a very good fielder. Also, he played most of his career in the 1890s, when batting averages were the highest. I have him rated at 19.5 wins, just below the marginal range for a Hall of Famer. Still, he was a pretty good player, and I think this was largely forgotten because of his long tenure and success as a manager. Fain was a good fielder as well, but he is ineligible because he played only nine years. I have him at 19.5, just below the marginal level. Keller comes in slightly higher at 22.2. He was a true slugger. Both were limited by injuries, and although very good players, I would not put them in the Hall of Fameclass. Keller did get a few votes, but had only six percent in his best year. DL: The most notable newcomer to next year's writer's ballot is Tim Raines. Hall of Famer? PP: I certainly rate Raines as a Hall of Famer, but I think he is not considered that way by many writers. I suspect he will fall short of election in his first try. Of those eligible for the next election, I have the following ratings:
1. Tim Raines (38.8) DL: Of the holdovers, who would you like to see voted in next year? PP: I think Gossage definitely deserves it. Bruce Sutter was a pretty good reliever, but I have him rated at 17.9 wins, only a little better than half of Gossage's mark. McGwire has borne the brunt of the steroid backlash, although I don't believe he did anything illegal. Hopefully he will get past that. He was hitting home runs in high school and did not put on that much weight during his career, unlike Canseco and Sosa. Blyleven and Trammell also are deserving. DL: Among those eligible beginning in 2009 is Rickey Henderson, who is a shooin. What about the other 2009, 2010, and 2011 firsttimers? PP: Henderson is an all time great. I have him rated 17^{th} all time, about the same as Mantle and Gehrig. Here is my list:
1. Rickey Henderson (72.1) eligible in 2009 Martinez has the problem of being a DH only, but 45 wins is well above the limit, so I think he is good. Palmeiro I don't think will ever be elected. I don't take the factors involving him into account in my system, so I will just let it stand at that. Walker is adjusted for Colorado, so he is legitimate. Bagwell, Larkin, and Alomar are good candidates. Kevin Brown had a tough time at the end, so that might affect his chances, but he was good for a long time. DL: What are your thoughts on Bernie Williams' Hall of Fame credentials? PP: I have Bernie at 25.4, right in the middle of the marginal range. He could go either way. DL: When Frank Thomas, Jim Thome, Jason Giambi, and Todd Helton are someday eligible, how will their careers be looked at? PP: Among active players, there are quite a few who will definitely get in. These are ratings through 2006:
1. Barry Bonds (124.7) All these guys are pretty near the end of their careers and are unlikely to add much more to their ranking, except of course for ARod. Only 31 years old, he could end up ahead of all of them. Of those lower down who still have a way to go, I would rate Pujols as most likely to end up the highest; he could reach ARod territory. Those likely to make 40 wins are:
1. Manny Ramirez (39.5) age 35 I don't think Thome (35.9, age 36) and Helton (31.3, age 33) will make 40, but they still could be good candidates. Jim Edmonds (35.6, age 37) is also in this group, along with Jeff Kent (36.7, age 39), John Smoltz (36.5, age 40), Mike Mussina (34.2, age 38, Craig Biggio (32.9, age 41), and Curt Schilling (30.5, age 40). Jason Giambi at 24.1, age 36, is a bit below. DL: How do you rate Ichiro Suzuki?
PP: I don't think Ichiro will make the Hall of Fame unless his Japanese stats are included. He is a good solid player, but his current lifetime rating is only 14.5 after six years at age 33 and he has declined for the past two years. I would be surprised if he passes 20 wins lifetime. And of course, he might not make 10 years. If there was an international Hall of Fame, he would be a good candidate. 
0 comments have been left for this article. (Click to hide comments) BP Comment Quick Links Not a subscriber? Sign up today!
