You Are Here > Baseball-Reference.com > Blog >

SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all B-R content, so you can quickly and easily find the content you want.

Also, our existing B-R blog rss feed will be redirected to the new site's feed.

Baseball-Reference.com ยป Sports Reference

For more from Andy and the gang, check out their new site High Heat Stats.

B-R.com’s WAR Replacement Level is .320 (52-110)

Posted by Sean Forman on September 13, 2010

I've seen some discussion here and elsewhere curious about where replacement level is set. We set it at a .320 W-L percentage. So a team with 0.0 WAR should be around a .320 team. And a team with 25 WAR should be 25 wins over .320.

Here are the 2010 teams so far this year.

This is through Sunday night. The columns are wins, losses, WAR, W_war (or number of expected wins based on WAR, the error in those numbers and then the Pythag win-loss.

WAR is going to be off more than pythag because it includes things like strength of schedule for pitchers and also includes a different replacement adjustment for the AL and NL because the AL is a stronger league by a good margin.

| team_ID | W    | L    | WAR | W_war | error | W_pyth |
+---------+------+------+-----+-------+-------+--------+
| NYY     |   87 |   56 |  44 |  89.8 |   2.8 |   88.0 |
| MIN     |   85 |   58 |  43 |  88.8 |   3.8 |   84.2 |
| TBR     |   86 |   56 |  40 |  85.4 |   0.6 |   85.6 |
| BOS     |   79 |   64 |  38 |  83.8 |   4.8 |   77.3 |
| ATL     |   82 |   62 |  37 |  83.1 |   1.1 |   84.5 |
| TEX     |   80 |   63 |  35 |  80.8 |   0.8 |   79.8 |
| SDP     |   80 |   62 |  35 |  80.4 |   0.4 |   82.7 |
| CHW     |   79 |   64 |  35 |  80.8 |   1.8 |   78.6 |
| TOR     |   73 |   70 |  34 |  79.8 |   6.8 |   73.5 |
| SFG     |   81 |   63 |  34 |  80.1 |   0.9 |   81.3 |
| PHI     |   83 |   61 |  33 |  79.1 |   3.9 |   80.6 |
| CIN     |   81 |   62 |  32 |  77.8 |   3.2 |   79.8 |
| DET     |   72 |   72 |  32 |  78.1 |   6.1 |   72.4 |
| COL     |   79 |   64 |  31 |  76.8 |   2.2 |   78.5 |
| OAK     |   71 |   71 |  31 |  76.4 |   5.4 |   74.5 |
| STL     |   74 |   67 |  31 |  76.1 |   2.1 |   79.3 |
| NYM     |   70 |   73 |  25 |  70.8 |   0.8 |   71.5 |
| FLA     |   73 |   69 |  25 |  70.4 |   2.6 |   74.4 |
| LAD     |   71 |   73 |  23 |  69.1 |   1.9 |   71.6 |
| LAA     |   70 |   73 |  21 |  66.8 |   3.2 |   68.3 |
| CLE     |   58 |   85 |  19 |  64.8 |   6.8 |   59.7 |
| WSN     |   60 |   83 |  18 |  63.8 |   3.8 |   64.6 |
| MIL     |   66 |   76 |  17 |  62.4 |   3.6 |   64.1 |
| KCR     |   58 |   84 |  17 |  62.4 |   4.4 |   54.6 |
| BAL     |   55 |   88 |  16 |  61.8 |   6.8 |   53.3 |
| CHC     |   62 |   81 |  15 |  60.8 |   1.2 |   61.8 |
| SEA     |   55 |   88 |  15 |  60.8 |   5.8 |   53.6 |
| ARI     |   57 |   86 |  14 |  59.8 |   2.8 |   60.8 |
| HOU     |   68 |   75 |  13 |  58.8 |   9.2 |   61.3 |
| PIT     |   48 |   94 |  -1 |  44.4 |   3.6 |   43.9 |
+---------+------+------+-----+-------+-------+--------+

This entry was posted on Monday, September 13th, 2010 at 9:52 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

33 Responses to “B-R.com’s WAR Replacement Level is .320 (52-110)”

  1. Johnny Twisto Says:

    Sean, I don't know what recent measurements you've actually done (or seen)....is the AL still quite superior to the NL? My impression is that the gap had closed compared to the past few seasons. I base that, I suppose, on the interleague record as well as my subjective impression that more talent had moved from the AL to NL than vice versa of late.

  2. It isn't a huge difference on the order of four runs difference between a starter in the AL and the NL.

    Also, the AL is still likely better.

    Interleague Play:
    AL leads NL 134-118 (1168 runs to 1098)

  3. How does Pit have a -1 WAR if they are better then .320 winning percentage? Or is that what you meant by the last paragraph?

  4. One thing I've noticed that seems a bit shaky with this version of "WAR" are the fielding numbers that come out of Total Zone. For Instance.

    Ryan Zimmerman is +2 in Fielding Runs while Mike Hessman is +3. How the heck is that even possible???? Hessman has only played about 40 innings of Third and 40 innings of First yet TZ would have us believe he's actually been a better defensive player than Ryan Zimmerman with 1000+ innings??

    Fan Graphs version of WAR has Zimmerman as one the best defensive players in baseball with a +15.9 fielding runs. This version of Zimmerman as one of the best defensive players in baseball would tend to go along with the consensus of him as a player. Overall Fangraphs "WAR" see's Zimmerman as the best player in the N.L.

  5. I still don't understand positional scarcity. I mean, I understand the CONCEPT of it, but how it's calculated remains a complete and utter mystery to me, especially since 2 guys who play the same position seem to have completely different numbers in that column quite often. And, for me, I personally put very little stock in most defensive metrics, and I've seen very little to change my mind. Until then, I'm just going to throw that column out in my mind. Also, yeah, saying the AL is significantly stronger reeks of AL-bias and sounds EXACTLY like the SEC claiming that the SEC is the strongest conference. (Spoken like a true NL and Pac-10 fan)

  6. I know this is out of place in a thread dedicated to WAR lovers, but I can't stand the stat.

    Maybe someone can change my mind. Just explain this one.

    1968. Denny McClain goes 31-6, era below 2, only unanimous MVP and Cy Young EVER. And yet, his 5.90 WAR does not rate in the top 500 individual seasons. NOT EVEN IN THE TOP 500? WTF!

  7. I knew a team with a total WAR of 0.0 would not be too good, but I never would have guessed it would be that bad.

  8. Barkfart/6 - It could be that pitcher W-L and award voting has no place in a metric like WAR. There's a lot of relevant numbers that were ignored, but that would keep McClain's numbers down.

    First: The Tigers scored more runs than anybody that year by a wide margin. League OPS was .637 that year. To compare, this year Felix Hernandez has a .599 OPS against. CC Sabathia is .666. Cliff Lee is .619. In other words, the AVERAGE AL hitter in 1968 performed about as well as a hitter does against Cliff Lee this year. There was no offense in the American League in 1968.

    League ERA in 1968 was 2.98. That means McClain's 1.96 ERA was a full run below average. That's certainly very good (as the 5.9 WAR suggests), but he just wasn't as far from average as a handful of pitchers in this year alone. When the ERA of the whole league is a hair under 3, one guy's ERA being a hair under 2 is a lot less impressive.

    I actually don't see how anybody could make a legitimate argument that McClain's WAR should be higher.

  9. Hessman has only played about 40 innings of Third and 40 innings of First yet TZ would have us believe he's actually been a better defensive player than Ryan Zimmerman with 1000+ innings??

    John, all I can say there is that the methodology is very consistent in how it treats players. It measures ground balls and where they are fielded on the field, the handedness of the batter and a park effect based on the number of ground balls converted into outs (think astroturf vs. grass). Without seeing every one Zimmerman and Hessman's plays it would be impossible to know whether the stat "made sense or not". Also the stat is against an average player, so if Hessman has been very good and Zimmerman above average their relative ranking would make complete sense.

    I still don't understand positional scarcity. I mean, I understand the CONCEPT of it, but how it's calculated remains a complete and utter mystery to me, especially since 2 guys who play the same position seem to have completely different numbers in that column quite often. And, for me, I personally put very little stock in most defensive metrics, and I've seen very little to change my mind. Until then, I'm just going to throw that column out in my mind. Also, yeah, saying the AL is significantly stronger reeks of AL-bias and sounds EXACTLY like the SEC claiming that the SEC is the strongest conference. (Spoken like a true NL and Pac-10 fan)

    The first stop for understanding positional scarcity is the league batting splits by position adjusted to 650 PA's. The system is set up (this is a crude simplification) so that if you add the average batting contributions at each position (above or below average) and the positional adjustment for each position you'll get 0 runs (or an average player).

    Can you show me two guys who play the same position and have wildly different #'s?

    Regarding AL vs. NL. First the adjustment is small. The average player in the AL (over the course of a season) is 22 runs above replacement and the average NL player is set at 18 runs. (or the difference is 4 runs or .4 wins.)

    Here is my justification for that adjustment. Anecdotally, the AL has won more World Series and also All-Star Games, but they have dominated interleague play.

    | year_game | W    | L    | RS   | RA   |
    +-----------+------+------+------+------+
    |      2010 |  134 |  118 | 1168 | 1098 | 
    |      2009 |  138 |  114 | 1206 | 1061 | 
    |      2008 |  149 |  103 | 1249 | 1014 | 
    |      2007 |  137 |  115 | 1352 | 1172 | 
    |      2006 |  154 |   98 | 1336 | 1115 | 
    |      2005 |  136 |  116 | 1230 | 1056 | 
    

    That's not a fluke and it is pretty clear the AL is playing better ball.

    Over that time period, the Royals are 58-50 in interleague play. Here are all 30 teams interleague record for the last six years. AL teams have nine of the top ten records and NL teams have nine of the ten worst records. The worst AL record, Cleveland at 48-60, is as good or better than that of ten National League Teams. If you want to see the strength of the SEC, I would urge you to visit this page on our College Football Site and sort by SRS.

  10. Lawrence Azrin Says:

    If an entire team of replacement-level players would have a record of about 52-110, how does one explain the 2003 Tigers, who went 43-119, nine games WORSE than "replacement level"?
    Possible explanations:
    - bad management, both on-field and player personnel choices?
    - random bad luck
    - one of those things

    I don't disagree with where replacement level is set, I am just wondering how one explains the 2003 Tigers.

    Also, don't there seem to be more really bad teams recently? From the article on 100+/loss teams since 1973, I counted a third of them being from the 2000's, and more of the REALLY crappy teams seemed to be from the 2000's, though I cannot quantify that as well. Any speculation?

  11. Lawrence, I would say all of the above. Extreme performances are by their very nature flukes. To be the worst ever at something you need to be really bad, make bad decisions and add a lot of bad luck to that as well. Just like a team that wins 110 games has to have everything go their way.

    Actually the number of 100L teams (actually < = .383 W-L record) has been pretty stable, 1990's excluded.

    | decade | count(*) | perc_100_L |
    +--------+----------+------------+
    | 1920's |      160 |       11.3 | 
    | 1930's |      160 |       15.0 | 
    | 1940's |      160 |        8.8 | 
    | 1950's |      160 |        9.4 | 
    | 1960's |      198 |        9.6 | 
    | 1970's |      246 |        6.9 | 
    | 1980's |      260 |        5.8 | 
    | 1990's |      278 |        1.8 | 
    | 2000's |      300 |        5.0 | 
    | 2010's |       30 |        6.7 | 
    

  12. Lawrence Azrin Says:

    #11/ Sean Forman Says: "Actually the number of 100L teams (actually < = .383 W-L record) has been pretty stable, 1990's excluded.

    | decade | count(*) | perc_100_L |
    +--------+----------+------------+
    | 1920's | 160 | 11.3 |
    | 1930's | 160 | 15.0 |
    | 1940's | 160 | 8.8 |
    | 1950's | 160 | 9.4 |
    | 1960's | 198 | 9.6 |
    | 1970's | 246 | 6.9 |
    | 1980's | 260 | 5.8 |
    | 1990's | 278 | 1.8 |
    | 2000's | 300 | 5.0 |
    | 2010's | 30 | 6.7 |"

    It's a great honor that you replied too my post, thank you Sean. Let me be the empteenth person here to thank you for the greatest website ever created (and the most enjoyable distraction from work).

    Looking at your table, it looks like I was fooled by the extremely low number of bad teams in the 1990's (1.8%), and accepted that as the norm, along with the truism that there's hardly any REALLY bad teams anymore, because of a higher level of competitive balance (unlike earlier in the 20th century). As the 2003 Tigers and 2004 Diamondbacks (three years after winning the WS!!)prove, historically bad teams can happen anytime.

    I've always had a perverse fascination with extremely bad teams, I mean I understand why an expansion team is terrible the first few years, but it's harder to understand how a team that's been around a century and had many outstanding years can be so awful...

  13. Johnny Twisto Says:

    I still don't understand positional scarcity. I mean, I understand the CONCEPT of it, but how it's calculated remains a complete and utter mystery to me, especially since 2 guys who play the same position seem to have completely different numbers in that column quite often.

    My guess is you are looking at guys with different playing time. A full-time SS might get +7 runs for position, but a September call-up doesn't get a +7 in that column as soon as they play their first couple innings there. It's going to be prorated based on PT.

  14. JD @8,

    The whole point of era adjustments are so we can measure 1930 pitchers vs. 1968 pitchers.

    When we do this, we are making a huge assumption that the same pitcher at his absolute peak, i.e Denny McLain in 1968, would definitely be PROPORTIONALLY better (or worse in this case) in another era/year. Meaning if you thake that 1968 McLain and put him in another season he would always have an ERA of about 2/3 the league average, as opposed to the idea that he might have a 1.96 ERA in no matter what year he pitched. Niether concept can possibly be proven, but we have so accepted the Adusted ERA value in our statistical analysis that we sometimes think of it as a proven fact.

    Therefore, in a year when the overall offense is WAY down (1968), a pitcher's adjusted numbers are devalued even more.

    I believe if one opens their mind to the idea that the peak McLain could give up the same hits/outs combination in the fictitious matchups with say 1930 batters that would lead to a 1.96 ERA (as opposed to the adjusted ERA - since it is not like batters were just given hits/runs in higher offensive eras - they still had to do it), then you could argue that Denny's WAR number is low due to possible over-adjustment for the 1968 pitchers (which corresponds to McLains great season).

    Before someone says that the adjusted stats is the best way we have to make up WAR, that is true, but we need to realize the basic assumptions made and how it does not lend to comments like "I actually don't see how anybody could make a legitimate argument that McClain's WAR should be higher". If you can prove without any doubt (meaning no assumptions) that the 1968 adjustment is perfect, then you are correct, but otherwise, not so much. We need to remember how conceptual these stats actually are before we make broad truth statements about anyone's perceived value.

  15. When we do this, we are making a huge assumption that the same pitcher at his absolute peak, i.e Denny McLain in 1968, would definitely be PROPORTIONALLY better (or worse in this case) in another era/year. Meaning if you thake that 1968 McLain and put him in another season he would always have an ERA of about 2/3 the league average, as opposed to the idea that he might have a 1.96 ERA in no matter what year he pitched. Niether concept can possibly be proven, but we have so accepted the Adusted ERA value in our statistical analysis that we sometimes think of it as a proven fact.

    I'm not sure I understand your point, but my answer is, "of course that is what we do." I don't think these stats try to answer, "What would Denny McLain have pitched like in 1930?" Of course, 1968 McLain would have dominated even moreso in 1930. As in almost every sport, the quality of play improves over time in baseball.

    The question I'm trying to answer is what is the value of McLain in 1968 to his team's winning and losing and how does that compare to the value of Grove in 1930 and his effect on his team's winning and losing. What Grove would do in 1968 is really a meaningless unknowable question, IMO.

  16. Here's a question though . . . assuming a 4.5 R/G league, a team with a .320 offense - meaning the team would play .320 with average pitching; and a .320 defense - meaning a team would play .320 with average hitting, would actually play .178 ball (29-133).

    To get the team to .320, you'd have a .407 offense and a .407 defense. In the 4.5 R/G environment, 3.661 R/G is .407 team with an average pitching staff, and 5.467 R/G is a .407 team with an average offense. 3.661 R/G and 5.467 R/A is a .320 overall team.

    How you define what .320 is makes a huge difference.

  17. Luck and Good fortune had a lot to do with McLain's '68 season.

    Out of the 84 pitchers that qualified, McLain led the Major League in Run Support per game with a 5.2. He was also the only major league starting pitcher to get more than 5 runs per game start in 1968.

    Also, the '68 Tigers were a pretty good defensive team.

    Sean,

    Great site by the way. Any chance you could set up something in the leader-board the separates the "Rate Stats" by time period? Something like OPS+ post 1901 would give a clearer picture of who were really the best hitters of all-time. As it is Dave Orr, Pete Browning, and Benny Kauff are among the the all-time leaders in OPS+. And the Minimums seem kind of low 3000 P.A. for example.

  18. John Q,

    That is what the Play Index Season Finder is for. :)

  19. I still don't understand how one can definitively make the argument that one league is stronger than the other.

    The AL may have won more World Series of late, but those results are skewed by (a) The AL winning the All Star Game more times since it started deciding home field advantage and (b) Half of those WOrld Series' being won by the Yankees and/or Red Sox, who have more money to spend and are on average stronger than any other team in either league.

    Note the Yankees and Red Sox being strong doesn't make the whole league really strong. It just makes 2 of the 14 teams disproportionately strong compared to the others.

    The All Star game happens once per year, and as far as I'm concerned who wins that game doesn't really mean anything. Starting pitchers are being used for 1 inning at a time. Players are batting in strange spots in the lineup. It isnt exactly an intense, competitive atmosphere. And of course, it's one game.

    I think the interleague record provides the best evidence for the AL being stronger, but even that isn't conclusive--after all, the two leagues play with different rules, and thus whenever teams from the two leagues play each other one must either lose its DH or add a DH to its lineup. Of course, the addition or subtraction of a DH changes the entire dynamic of a batting order, and it may be that for whatever reason not having a regular DH is more of a burden for NL teams in AL parks than having to go without a DH is for AL teams in NL parks. I would like to see some interleague home/road splits for each league.

    Of course, the fact that more runs are scored in the AL than are scored in the NL means nothing other than that the AL is a more offense-oriented league (mostly on account of having a DH). It doesn't make the AL stronger, it just means the balance of offense to pitching is a little different.


  20. Of course, the fact that more runs are scored in the AL than are scored in the NL means nothing other than that the AL is a more offense-oriented league (mostly on account of having a DH). It doesn't make the AL stronger, it just means the balance of offense to pitching is a little different.

    The runs scored and allowed are just in the interleague games.

    Malcolm, click the link in my comment that goes to the interleague records. It isn't just the Yankees and Red Sox. Nine of the top ten records in interleague play the last six years belong to American League teams. NINE OF THE TOP TEN. The worst AL team record (Cleveland's) is as good or better than 2/3rds of the teams in the NL.

  21. Also, the American League has a .505 winning percentage In the NL Parks, so even though they have the road-field disadvantage they still win more games on the road.

  22. How do you figure W_war?

    If its not a trade a secret to explain.

  23. W_war is just adding up all of the WAR of the player's on the team.

  24. Barkfart, let me show you how McLain's WAR was calculated.

    He allowed 86 runs. (This is the actual total, including unearned runs). Because 1968 was such an extreme pitchers year, the replacement level for a pitcher was very low, just under 4 runs/9 innings. For Denny's 336 innings this gives us a replacement level of 147 runs, in Tiger Stadium. The Tiger's were a fine defensive team that year, 13 runs better tan average over 336 innings. This puts replacement level for a Tiger's pitcher that year at 134 runs. Denny was 48 runs better than that. (48 RAR.) Because run scoring that year was so low it took fewer runs than the historic norm to add a win, closer to 8 than the more normal 10. (In the present high scoring era, it is a bit above 10). Thus 48 RAR becomes 5.9 WAR.

    Compare with Gibson that year. He allowed 49! runs, his replacement level was also 134 runs. So he was 75 RAR. I guess the NL was even lower scoring than the AL because this converts to 11.9 WAR.

  25. Kds- thanks so much for the explanation. I really appreciate it. I guess old-school gut-level guys like me are gonna have a hard time coming over to the WAR side.

    To say that the guy could've pitched substantially better than he did is hard to take.

    Ask the people that played with him and they'll tell you it was one of the greatest season's ever. WAR says it doesn't even crack the top 500.

    I think WAR overcompensates Steve Carlton for his 27 win season (and a GREAT season it was!) because he played on a bad team, and then penalizes McClain for playing on such a great one.

    thanks again

  26. Michael E Sullivan Says:

    "then you could argue that Denny's WAR number is low due to possible over-adjustment for the 1968 pitchers (which corresponds to McLains great season)."

    You could argue that, but it doesn't make a lot of sense. The fact is that McLain's 1968 season, while excellent, was not dramatically or unusually excellent compared to other top pitchers that year in any metric except number of wins. McLain's 1.96 ERA looks spectacular to 2010 or 1930 eyes, but he was only fourth best in the AL in 1968, and none of the guys in front of him had fewer than 250 IP. Tommy John was right behind him with 1.98, and another guy had a 2.05, and none of these guys were relievers, or playing short seasons. You could argue that McLain's 336 IP gives him a leg up on the guys who finished right around him in ERA, but Tiant managed a 1.60 that year, on 258 innings. Looking over in the NL, Bob Gibson managed a 1.12 ERA that year in 300+ IP. That's what a dominant season looks like, and his 258 ERA+ shows it. McLain was fifth in MLB ERA+ that year, with an excellent, but hardly earth-shattering 154, the kind of mark that will get you on the leaderboard most seasons, but rarely claim the top spot.

    So having a sub 2.00 ERA in 1968, while still a mark of excellent pitching, was not the kind of really dramatic outperformance even of other top pitchers that you see with, for instance, Pedro's 1.74ERA in 2000. Look at the leaderboard. The next closest pitcher in the major leagues was Kevin Brown's 2.58 in the NL. There were a total of 5 guys in the majors that year with 3.00 or less, and everybody but Pedro was in the NL, where the avg ERA was .28 runs lower. The next best AL pitcher was Roger Clemens at coming in at 3.70, more than *double* Pedro's ERA with fewer IP. He was followed by Mussina at 3.77 and two other guys below 4. Only 5 AL pitchers had ERA qualifying seasons under 4.00, and Pedro came up with 1.74. *That* is dominance. It's also one of the best pitching seasons in baseball history, and the guy only won 18 games that year.

    McLain's 1.96 was a great year, for sure, but it is memorable primarily because he played so many innings on such a potent offensive team with enough luck to crack the arbitrary 30 win barrier, not because his pitching was so much better than other pitchers at the top of the heap.

  27. Lawrence Azrin Says:

    #26/ "Michael E Sullivan Says:
    You could argue that, but it doesn't make a lot of sense. The fact is that McLain's 1968 season, while excellent, was not dramatically or unusually excellent compared to other top pitchers that year in any metric except number of wins. You could argue that McLain's 336 IP gives him a leg up on the guys who finished right around him in ERA, but Tiant managed a 1.60 that year, on 258 innings. Looking over in the NL, Bob Gibson managed a 1.12 ERA that year in 300+ IP. That's what a dominant season looks like, and his 258 ERA+ shows it."

    You could also argue that Tiant's 1968 was just as good or better than McLain, but he didn't have the same offense behind him. You could also argue that both Tiant and Bob Gibson (who had roughly the same W/L records) might have also won close to 30 games, if they had the same run support as McLain. As Micheal says, McLain was great in '68, but not historically great like Gibson was.

  28. Indis W/L

    I also have a process to convert WAR into Individualized W/L records. Link above is for 1968. Bob Gibson was the clear leader of that class of players.

  29. David in Toledo Says:

    Sean, I really appreciate your taking up this subject.

    Does this mean that a replacement-level team pre-expansion (154 games) would theoretically have a record of 49-105?

    I'm taking the 1954 Indians as an example. Adding 27.5 team batting WAR and 23.6 team pitching WAR would raise that theoretical record to 100-54, correct? (Actually, to near 101-53, since 32% of 154 is 49.3.) Then fielding. . . I assume team WAR for fielding isn't available. But it appears all the Cleveland regulars were above average in their defensive numbers for 1954 except Al Rosen. So that nudges the record upward, into the range (none of this can be expected to equal the exact playing record) of 104-50.

    And if that isn't 111-43, well, neither is Cleveland's 1954 pythagorean (it's 104-50, same as the Yankees'). The Indians would appear to have had better luck than the Yankees, and it certainly helped them to win 1-run games that they had the pitching staff and particularly the bullpen that they did.

    Am I correct that the use of historical WAR would predict the 1954 Indians to finish in the range of slightly/somewhat better than 101-53, but not actually as high as the 111-43 on the record books?

  30. David,

    I would expect it to be closer to the Pythag record than the actual record, though they all should be in the same ballpark.

  31. Johnny Twisto Says:

    David, the "batting WAR" you cited is actually position-player WAR, which does already include the fielding.

  32. Johnny Twisto Says:

    I think WAR ... penalizes McClain for playing on such a great one.

    It's not really penalizing him. It is saying that, as best we can extract the individual performance from that of his team, his individual performance was not quite as great as it appears because the support of his team was also great. He had a strong defense, which affects his runs allowed. He had a strong offense, which affects his win total.

  33. David in Toledo Says:

    Thanks to you both for helping with my understanding of 1954 Indians WAR. It would appear that the expected record of that team, based on the cumulative WAR of all its individual players, would be 100-54 or 101-53. The actual record was 111-43, or about 10% higher. Is a 10% deviation at the extreme of what one would find when summing individual WAR and comparing the total to an actual team record? Have you run a program to find out what the typical difference is between these team WAR sums and actual records?

    I figure putting Minnie Minoso in Cleveland's outfield for 1954 instead of Dave Philley would have given the 1954 Cleveland Hypotheticals -- using WAR -- a WAR record of about 109-45. Still not 111-43, but with no possible doubt that the Yankees could somehow have grabbed yet another pennant.