The Book Blog Comments

Latest Articles in this Channel:

11/30/11--22:09: Forecaster’s Challenge: Final 2011 Results (16:13:31-05:00) (chan 1600204)

By Tangotiger

I ran four competitions, three unofficial, and one official. I’ll run them all down.  I’m going to list the results of all the pro forecasters who finished ahead of Marcel.  For those that finished below Marcel, I will list them in alphabetical order. 


FAN_ID: a unique number for internal purposes

POINTS_CT: the average points for each team.  The league leader in Fantasy points has around the same number of points as the league leader in RBIs does.  That’s the kind of scale you should think of. 

WINS_CT: the number of leagues won.

VALUE_CT: the number of draft points won in the league.  You get 11 points for a win, 5 for a 2nd place finish, 3, 2, 1 for 3rd through 5th.  C

First, the unofficial competitions.

Setup 1: All Pros
All 22 Pros in the same league, 100 leagues.

We see that the winner is the Consensus pick, which is the consensus of the other 21 forecasters.  Rotoworld was a close second (and Rotoworld also finished in 2nd place in the inaugural competition two years ago).  Steamer, a forecasting system created by students of Jared Cross did well. 

Other highlights are the Fangraphs Community being average, and Marcel not doing well at all.

FAN_ID    POINTS_CT    WINS_CT    VALUE_CT    FAN_TX
299    1474    39    597    Consensus
122    1462    31    526    RotoWorld
118    1319    15    256    Steamer
113    1354    4    194    FEIN
116    1323    3    141    KFFL
102    1307    2    123    Ask Rotoman
112    1280    2    105    Mike Podhorzer_FantasyPros911
106    1312    1    95    CAIRO
133    1197    1    62    Rotochamp
120    1136    1    40    Razzball
135    1245    0    32    Pat Senechal
125    1237    0    12    BigScoreSports
131    1050    1    11    Fangraphs Community
115    1224    0    7    John Eric Hanson
132    1165    0    4    Fantistics
126    1090    0    0    Bloomberg Sports
217    1089    0    0    Marcel

Baseball Info Solutions
Future of Fantasy
Geoff Buchan
PECOTA
Statspeakblog

Setup 2: Head-to-Head Pros
Two-person league, each Pro facing off against one other Pro, 42 leagues each Pro

Rotoworld won 41 of its 42 leagues, while the Consensis was close behind.  Marcel has a slightly above-average showing, winning 24 of 42.  The Fangraphs Community was right behind Marcel.

FAN_ID    POINTS_CT    WINS_CT    FAN_TX
122    14175    41    RotoWorld
299    14206    39    Consensus
113    13710    37    FEIN
118    14053    36    Steamer
112    14040    34    Mike Podhorzer_FantasyPros911
106    13683    30    CAIRO
116    13493    28    KFFL
120    13587    28    Razzball
132    13544    28    Fantistics
217    13311    24    Marcel

Ask Rotoman
Baseball Info Solutions
BigScoreSports
Bloomberg Sports
Fangraphs Community
Future of Fantasy
Geoff Buchan
John Eric Hanson
PECOTA
Pat Senechal
Rotochamp
Statspeakblog

Setup 3: 1 Pros v 21 Random Joess
Each Pro faces off against Random Joes, whereby Random Joes created based off the Consensus of Pros (Value for each player for each Random Joe is within +/-5$ from the Consensus for 95% of the players, with the other 5% of players effectively removed from the pool for the Joes); 22 leagues

Marcel does pretty well here.  KFFL takes the top honors.  So far, we’ve seen KFFL, CAIRO, FEIN, and Mike Podhorzer all finish ahead of Marcel in all three unofficial competitions.

FAN_ID    POINTS_CT    WINS_CT    VALUE_CT    FAN_TX
116    1552    16    189    KFFL
106    1501    12    170    CAIRO
112    1498    15    169    Mike Podhorzer_FantasyPros911
113    1476    12    163    FEIN
135    1463    11    146    Pat Senechal
217    1460    11    134    Marcel

Ask Rotoman
Baseball Info Solutions
BigScoreSports
Bloomberg Sports
Consensus
Fangraphs Community
Fantistics
Future of Fantasy
Geoff Buchan
John Eric Hanson
PECOTA
Razzball
RotoWorld
Rotochamp
Statspeakblog
Steamer

***

Finally, the official competition.

Setup 4: 1 Pros v 21 Random Fangraphs Readers
Large collection of Fangraphs readers were pooled in various fashions to create 21 overall draft lists for each league; 22 leagues

In this competition, each Pro forecaster is put in his own league against 21 average forecasters (Joe).  In order to make the apples-to-apples comparison, the Pro faced off against the same 21 average forecasters, from the same draft spot.  So, when Marcel drafts first against Joe1 through Joe21 in the Marcel Draft1 league, then Rotoworld also drafted first against Joe1 through Joe21 in the Rotoworld Draft1 league.  Each of the 22 Pro forecasters followed the same pattern.  In addition to that, Marcel then drafted second against 21 new Joes (Joe22 through Joe42).  And so on, and so on, and so on.

This is I think the one setup that most mimics reality. This model, I believe, should represent a reasonable group of average Joes that you would find in a Fantasy League.

And the official winner is Ask Rotoman.  Congratulations Ask Rotoman!  It barely edged out Consensus by a single point and a single win.  Steamer and Rotoworld also had a strong showing.  The Fangraphs Community did about average, and Marcel did not do well.

FAN_ID    POINTS_CT    WINS_CT    VALUE_CT    FAN_TX
102    1502    14    176    Ask Rotoman
299    1528    13    175    Consensus
118    1517    11    161    Steamer
122    1515    10    155    RotoWorld
112    1464    9    142    Mike Podhorzer_FantasyPros911
113    1472    9    129    FEIN
132    1441    8    120    Fantistics
131    1434    6    103    Fangraphs Community
116    1415    8    102    KFFL
106    1386    5    84    CAIRO
135    1385    2    66    Pat Senechal
120    1341    1    42    Razzball
115    1330    2    35    John Eric Hanson
133    1313    1    30    Rotochamp
126    1300    1    28    Bloomberg Sports
217    1291    0    23    Marcel

Baseball Info Solutions
BigScoreSports
Future of Fantasy
Geoff Buchan
PECOTA
Statspeakblog

Since there are 22 points allocated per league, and there are 22 participants per league, the average numbe of points is 1 per team.  And with 22 leagues for each Pro, that means the average number of points for a Pro is 22.  As you can see, most Pros were above the league average.  That’s because the league average included all those Joes, and those Joes did not do well.

This is actually a huge selling point for the Pros: almost *any* forecasting system is better than *no* forecasting system.

Conclusion

Unless you are very confident in your system, you would be foolish to stick to a single system.  Hanson won two years in a row, but finished middle of the pack this year.  Marcel did great last year as well, but not so well this year. But, everyone believes their system is an exception.  So, save yourself the trouble: let everyone else worry about their system, then just take the consensus of those.  You’ll do great.

Post-script:

I also turned everyone’s scores in each of the four competitions into a z-score (number of standard deviations from the mean).  I averaged the scores out, and here are the results.

We see that even though Ask Rotoman won the official competition, it was Consensus that was able to do the best across all four setups.  Given that it barely lost to Ask Rotoman in the official competition, it’s heartening to see Consensus (essentially the S&P500 index) do as well as it did. 

Marcel finished in the middle of the pack.  Think about that.  There are 20 other individual pro forecasters (excluding the Consensus), and we know they put more thought into their picks than Marcel did.  And yet Marcel was beaten by 10 of them and lost to 10 others.  This is the kind of result you’d expect if there was no such thing as talent at being a forecaster.

ALL    Pros    H2H    Joes    Readers    fan_id    fan_tx
1.55    3.00    1.43    0.17    1.62    299    Consensus
1.42    2.57    1.59    0.21    1.30    122    RotoWorld
1.01    0.94    1.19    0.49    1.40    118    Steamer
0.97    0.57    1.27    1.16    0.89    113    FEIN
0.86    0.03    1.03    1.29    1.09    112    Mike Podhorzer_FantasyPros911
0.74    0.25    0.56    1.72    0.46    116    KFFL
0.54    -0.03    0.71    1.31    0.17    106    CAIRO
0.43    0.14    -0.32    0.26    1.64    102    Ask Rotoman
0.22    -0.58    0.56    0.15    0.74    132    Fantistics
-0.07    -0.41    -0.56    0.79    -0.12    135    Pat Senechal
-0.08    -0.36    0.56    -0.02    -0.50    120    Razzball
-0.16    -0.60    0.24    0.54    -0.81    217    Marcel
-0.19    -0.54    0.16    -0.86    0.47    131    Fangraphs Community

Baseball Info Solutions
BigScoreSports
Bloomberg Sports
Future of Fantasy
Geoff Buchan
John Eric Hanson
PECOTA
Rotochamp
Statspeakblog




01/16/12--17:22: Clubhouse Confidential (17:01:48-05:00) (chan 1600204)

By Tangotiger

Brian Kenny is turning into a saberist’s best friend.  He has had Vince Gennaro, Joe Sheehan, Jay Jaffe, among others, and all have contributed positively.  Now, it’s Dave Cameron’s turn on the hot seat.




01/22/12--11:46: Fielding-Independent Perfect Games (15:11:31-05:00) (chan 1600204)

By Tangotiger

Nice idea (though it doesn’t seem he included hit batters).

To be more strict, I’d not only make it no HR, no BB, no HBP, but also so that you have a FIP of 0.00.  Since we have to add a constant of 3.2, that would simply mean if this term ends up at -3.2 or lower: 2*K/IP. So, if you throw 6 innings, that means 10 K. If you throw 7 innings, that means 11K.  8 innings means 13 K, and 9 innings means 15 K.

Under those conditions, how many Fielding-Independent Perfect Games have there been?





01/23/12--22:29: When athletes make political statements (02:05:21-05:00) (chan 1600204)

By Tangotiger

Tim Thomas declines offer from Obama.

Back in the 1970s, when Quebec was riding its separation wave, there was huge pressure on the french players on the Montreal Canadiens to support the separatist party that eventually came into power.  From my memories of that time (I wouldn’t even turn 10, so take that for what it’s worth), all the players remained neutral.  They reasoned it wasn’t their place to exude their influence on a topic they didn’t earn a right to influence on.  It’s one thing if it’s Angelina Jolie, who uses her celebrity to further her non-entertainment passions.  It’s another for a player to be conscripted.

Where does Tim Thomas fall?  I have no idea.  I would only hope that he speaks from the heart.  But I agree with the author that it seems rather impolite, and intolerant, to refuse dinner with your country’s democratically-elected president.  What others consider an honor, he considers it rather lacking.  It’s acts like his that are divisive.

I haven’t given this any thought, so feel free to educate me.






01/23/12--22:29: Explaining NFL OT Rules (16:01:33-05:00) (chan 1600204)

By Tangotiger

I think the ref did a good job of explaining the OT rules.  I’m not sure if there’s a great way of explaining it.  He was pretty clear, and didn’t give too much information in each sentence.  He wasn’t terse, and he wasn’t verbose.  I liked it.

What wasn’t clear is what happens on a safety on the initial drive (reverse?).  Presumably, it should end the game, but maybe there is an exception there as well?




01/25/12--19:06: Rod Cross: Smooth patch on a baseball, to get a curve (22:56:13-05:00) (chan 1600204)

By Tangotiger

Looks like fascinating stuff.  I believe Rod Cross, Emeritus Professor at the University of Sydney, is a partner of Alan Nathan.

You can watch the whole thing, or start at the 2:30 minute mark (which is where I have it set).  Then there’s video of a pitch from a game.

Glove-slap: G.M.




01/25/12--19:06: Prince Fielder (22:12:27-05:00) (chan 1600204)

By Tangotiger

Prince Fielder: signed for 9 years, 214MM$; paying for 6 wins in 2012, dropping by 0.5 wins each year; paying for 5MM$ a year, increasing by 5% each year.




01/25/12--19:06: Can you walk off the island? (20:57:50-05:00) (chan 1600204)

By Tangotiger

I’d like to see the BB and K rate for all leagues around the world.

It turns out that you can, in fact, walk off the island. The league-wide walk rate is a robust 8.3% against a 17.9% K rate. In this tiny sample that means marginally more walks and fewer Ks than the big leagues in 2011.




01/26/12--17:13: AL v. NL in 2011 (20:47:03-05:00) (chan 1600204)

By mgl8@cox.net

It is generally accepted in the sabermetric community that the AL is a better league than the NL, at least for the last several years.  This is evidenced by the fact that the AL has a large advantage in IL games, although at least some of that edge could be something other than overall “talent”, although this is not likely and several people, including myself, have found little or no inherent advantage to the AL in IL games (e.g., the NL teams do not have any DH’s, so they have to juggle their lineup in AL parks, on the other hand, in NL parks, AL teams have to sit their DH’s or juggle their lineup, perhaps putting a bad defender - their DH - in the field, the AL pitchers typically are poorer hitters than the NL pitchers, etc.).


Since 2005 (an arbitrary beginning point mind you), the AL has a .555 WP in IL games.  That is a lot!  That suggests an average AL team would be a 90 win team in the NL.

My research in the past (do a search in this blog and I think I wrote a piece a while back on THT or BP) indicated that the edge has mostly been in pitching and in fact the NL may have had equal or better hitting than the AL in the last few years.  Keep in mind this has nothing to do with league rpg or ERA or the fact that the AL bats 9 real hitters and the NL only 8 (plus pinch hitters for the pitcher of course).  When we say that the AL is a better pitching team that means that if you took a pitcher from the AL and put him the NL, his rank among pitchers and thus his value, like WAR, would go up.  Same for batters.

So one way to quantify (within some degree of certainty - actually not a whole lot) any differences is to look at players who switch leagues from one year to the next.  If a batter switches from the NL to the AL and his batting lwts or WAR per x number of PA (or any other rate stat that is relative to league average) goes up then he is going to a league with less batting talent and vice versa. Again, you cannot use some non-league relative stat like OPS or ERA or FIP.

A few things about that though. One, you have to look at “switchers” from both leagues since it is possible that all players do better or worse for one year after they switch for various reasons.  Two, you have to be careful about regression.  Players who switch leagues are on the average below average players so their stats will naturally get better regardless of whether the new league is better or worse than the old one.  Three, the results only tell us about the relative difference between one league in one year and the other league in the next year.  You have to do some more gymnastics to compare leagues from the same year. For example, let’s say that in 2010 both leagues had the same talent on offense, but in 2011 the NL got better.  If a hitter moves from the AL in 2010 to the NL in 2011, his league-normalized rate stats will go down even though the leagues were the same in 2010 and we don’t know yet about 2011.  We only know from that example that the 2011 NL was better than the 2010 AL.

Oh, and you have to account for age of course when using this method.

Another method, and one I will present herein, is looking at IL games.  For that, you have to look at lots of things, since the teams that are fielded in IL games are not necessarily the ones that are fielded in non-IL games and of course we don’t really know how to compare the pool of parks in each league (although someone can try computing PF’s based on IL game data).

Here is the basic method:

Look at how AL batters do against NL pitching in NL parks and compare that to NL batters against NL pitching in the same (NL) parks.  If the AL batters do better than the NL batters in the same parks against the same pitchers, then they are the better hitting league, right?  Not so fast.  You have to control for the pool of pitchers and batters, especially the latter.  As I said, the teams that each league fields in IL and non-IL games are not the same.  AL DH’s are sometimes benched in IL games in NL Parks, and NL lineups can look completely different in AL parks. We also have to look at both sides.  As with the first method, it is possible that batters in IL games do poorly just because they are facing unfamiliar pitchers or playing in unfamiliar parks.  That might be true of pitchers as well.

Anyway, to compare leagues in pitching (and defense), we look at how NL pitchers do versus NL batters in NL parks and how AL pitchers do versus NL batters in NL parks.  Etc. (we look at all the permutations where we control for the opposition and the park).

Lastly, we can actually look at defense somewhat independently, since UZR is not really that park or league dependent.  Presumably a hard hit ground ball in the NL to location X has around the same chance of being fielded as one in the AL.

Interestingly, the AL had a huge (relatively speaking) advantage in UZR in 2011.  They had a total of +47 runs and the NL had -47 runs (remember that UZR is always zeroed out each year for BOTH leagues combined and the base line for the catch rates for each bucket is always 6 years of data from both leagues).  So the AL had an advantage on defense of around .04 runs per game, which is a .510 WP alone! In 2011, the AL won 52% of the IL games, so there isn’t much left, assuming that the 52% win rate is the true difference in talent between the leagues, which is unlikely (one SD by chance in 252 IL games is 3.2%, so we are 95% confident that the true difference, due to talent, is between 58% for the AL or 54% for the NL).  So really, we can pretty much ignore the 52%, although the fact that in prior recent years the AL dominated, changes the Bayesian a priori estimate (going into 2011, the chances of the AL being better is large).  But enough of that Bayesian nonsense!

Here is rest of the 2011 (other than above-mentioned defense - UZR):

Actually, I’ll redo the defense to look at the UZR of players actually on the field in IL and non-IL games, because we have to factor out the defense from the pitching in each bucket of games.

Let’s start with the NL batting at home versus AL pitchers (IL games).  After adjusting for the pool of batters and pitchers on the field (compared to the average pool of batters and pitchers for the whole year), the NL had a lwts (it doesn’t matter what out value I use, since I am going to use the same out value for all lwts calculations) per 500 PA of -6.9 runs.

NL batting at home versus NL pitchers is -.9 lwts, or -.07 rpg.  That suggests that the AL pitchers are 6 runs per 500 PA (.468 rpg) better than the NL pitchers. That is a lot. But…

Again, maybe the NL batters did worse against AL pitching only because they are not as familiar. We also have the better AL defense.  Plus we are only looking at half of our sample - pitching on the road. We have to look at pitching at home.

Now let’s look at the other half of our sample.

AL batting at home versus NL visitors: 2.7 lwts runs.  AL batting at home versus AL visitors: -1.9 runs.  A difference of 4.6 runs.  The difference in NL parks was 6 runs.  We have to take the average and assume the rest is due to unfamiliarity with the pitchers or random fluctuation, or both. So the difference in pitching (plus defense) is (6+4.6)/2, or 5.3 runs per 500 or .413 rpg.

Let’s factor out defense.  The AL defense in NL parks was .035 runs per 500 PA and the NL defense on the road in NL parks was -.313 runs per 500 PA, so the total difference in defense in NL parks was .348.  In AL parks, the AL road defense was .408 and the NL road defense in IL games was -.781 (ouch). The difference in defense in AL parks, was 1.189. Again, we take the average of 1.189 and .348, or .768 runs per 500.

So we subtract .768 from 5.3 and we get 4.53, or .353 rpg for the AL advantage in pitching in 2011!

The batting is easier since we don’t have to worry about factoring out defense.  I’ll skip the individual numbers and just give you the result.

The NL actually had a .1 runs per 500 PA advantage in batting! That is .008 rpg.

So final tally is:

Pitching

AL by .353

Batting

.008 NL

Defense

.038 AL

Total: .383 rpg AL, which is a 54.2% WP. Voila!




01/26/12--17:13: rWAR v fWAR?  No.  rWAR + fWAR. (19:56:13-05:00) (chan 1600204)

By Tangotiger

Moving posts from another thread here.




01/26/12--17:13: Academic ivory towers and gated databases (14:13:02-05:00) (chan 1600204)

By Tangotiger

Peer-reviewed journals: it’s been nice knowing you.

Step back and think about this picture. Universities that created this academic content for free must pay to read it. Step back even further. The public—which has indirectly funded this research with federal and state taxes that support our higher education system—has virtually no access to this material, since neighborhood libraries cannot afford to pay those subscription costs. Newspapers and think tanks, which could help extend research into the public sphere, are denied free access to the material. Faculty members are rightly bitter that their years of work reaches an audience of a handful, while every year, 150 million attempts to read JSTOR content are denied every year.

And, perhaps, the future has arrived.





01/27/12--21:26: Relocated Team Doctrine v Stationary Team Doctrine (01:06:23-05:00) (chan 1600204)

By Tangotiger

Great post:

Dale Hawerchuk was born in Toronto and played major junior in Montreal. He was drafted by Winnipeg, where he set records for nine years, and then traded to Buffalo. He skated briefly in St. Louis, ended his playing career in Philadelphia, and now coaches in Barrie. And yet it is only in an arena in Arizona that his number 10 hangs, retired, over the ice.
...
The solution to this is (like the Japanese answer to the paradox of Theseus’ ship) to assert that the essence of the team is in its form and its function rather than its substance. A team is not what it is but what it does, defined not by the specific people or buildings or shirts that it uses but by its position in a system of social relations. And in the case of a team, that position is inextricably and fundamentally linked to a place.


Now, I wouldn’t necessarily tie it to a place.  After all, if the Jets move from Long Island to NJ (or wherever they came from), that’s not a new thread.  What matters is the fan base.  Did the fan base follow the team, or not?  If so, link them.  If not, sever them.

So, you do this on a case-by-case basis, and thinking like an historian, and not trying to fit things into simple slots with simple rules.  Rules of thumb by definition don’t work all the time.

Did Jets fans follow the Coyotes?  Did they abandon the Coyotes when the new Jets arrived?

There’s no doubt that this applies to the Expos/Nationals:


Surely, a team that has to move is a black eye for the League, but a team that simply ceased to exist would be two black eyes, a broken nose, and a kick in the nuts. Because of this Doctrine, the NHL can say that it has not ‘lost’ a franchise since WWII killed off the New York Americans. Every other failed team has been bought or merged elsewhere, and every time that happens, the NHL manages to avoid taking direct responsibility for an unjustified or unstable overexpansion.

I don’t know how the Dodgers and Giants moving cross-country should be handled.  What we need is an understanding of their fan bases, and what happened after the Mets showed up.





01/27/12--21:26: Calvin & Hobbes (00:46:01-05:00) (chan 1600204)

By Tangotiger

Bill Watterson made the comics Hall of Fame on peak value. 

It’s clear when you look at things outside of baseball that there are different standards, no?  Had The Beatles only done Sgt Pepper, don’t you put them in the Hall of Fame?  Dr Richard Daystrom or Zefram Cochrane make it on one invention, no?  Da Vinci makes it for Mona Lisa?  After the 4th Wimbledon, don’t you immortalize Bjorn Borg?

So, apply the same thing to baseball.  Have different standards for best-3 years, best-5 years, best-7 years, best-10 years, best-15 years, best-career.  Obviously, you’ll have very few qualified for the best-3 years, and the more years, the more players qualify.  Bonds 1999-2003 guarantees enshrinement, regardless of what else he did. 

I read the entire Calvin & Hobbes collection to my boy last year.  Some really great stuff.






01/27/12--21:26: Who’s evaluating the 2011 forecasts this year? (15:07:38-05:00) (chan 1600204)

By Tangotiger

Anyone going to step up?  Anyone? The hard part is collecting all the data, and matching all the players.  If someone ELSE does all that hard work, I can step in and do the rest.

The test is pretty simple.
1. Calculate wOBA for every forecast, and for the actual.  I’ll do something simple like
numerator = 0.7*BB + 0.9*1B + 1.3*(2B+3B) + 2*HR
denominator = BB+AB

It really doesn’t matter much what you do here.  You just need something that focuses on the important stats, and make sure everyone forecasted those stats. 

2. Calculate each population mean, by weighting by actual PA (AB+BB).  For missing players, either give them the population mean (you HAVE to do this for Marcel, since by definition, Marcel has no missing players), or set the wOBA at 20 points below the rest of the population mean.

3. Recalculate new population mean (where applicable).

4. Baseline each player to a common mean (set to .330, but it doesn’t really matter what you set it to).  So, if the pop mean in #3 is .327, and you have a player forecasted for .377, his adjusted forecast is .380.

5. Calculate the difference for every player.

6. Present the average absolute difference, and the RMSE, and in both cases, weighted by the actual PA.

That’s it, that’s the basis.

Then you can do fun stuff, like splitting by career performances.  Guys with 1500 or more PA in the last 3 years, guys with fewer than 250 PA, guys who had a .380+ wOBA in the last three years, shortstops, etc, etc.  Look for whatever attribute of a player you want.  And compare the systems, and look for bias.

Bueller?




01/27/12--21:26: Do relievers today perform better because they have shorter outings? (00:13:03-05:00) (chan 1600204)

By Tangotiger

Dave is arguing that they do not:

On the other hand, strikeout rate has skyrocketed, increasing by 40% since 1982. This would seem to support the idea that relievers can be more effective in shorter stints, and that playing the match-ups can help prevent run scoring. However, there’s a problem with that theory – the strikeout rate of starting pitchers has gone up 41% during the same time frame. While strikeout rate has been raising at the same time that the modern bullpen has been evolving, this seems to be a case where correlation is not causation. If starters are seeing the same rise in strikeout rate, that points to a more fundamental shift among hitters – more sluggers swinging for the fences, the rise in acceptance of the strikeout as just another out among organizations – rather than a specific benefit being given to relievers from their new roles.

It would seem to me that this is evidence that they ARE performing better.  While the number of starters per team has remained fairly stable (after all, you have 162 starts per team, and pretty much everyone is on a 5-man or 5-day rotation, and they average 100 pitches per start), the number of relievers have skyrocketed.

What happens when you give more jobs to guys on the bubble?  Well, the overall average must go DOWN.  If for example the average team in 1982 used 10 different relievers, and the average team in 2002 used 15 different relievers, then those five extra relievers would have been in AAA in 1982.  The overall average in talent therefore would have gone down. But, we see the PERFORMANCE has remained relatively stable.  This would imply that the good relievers must pitch better in shorter outings.

Let me try to illustrate numerically.  I’ll just give some made-up numbers for talent:


Talent    Reliever#
150    1
120    2
100    3
90    4
85    5
80    6
75    7
70    8
65    9
60    10

The average of the top 5 relievers on a team would be 109.  But the average of the top 10 relievers on a team would be 89.

But, what if we add a new column that shows the performance level for each reliever, if we give them less to do:


Talent    Reliever#    Performance
150    1    170
120    2    140
100    3    120
90    4    110
85    5    105
80    6    100
75    7    95
70    8    90
65    9    85
60    10    80

Now, the performance level of the top 10 relievers is 109.

So, just as we know that the performance level of pitchers jumps substantially when pitching in relief than when pitching as a starter, it’s very possible that the performance level jumps somewhat when pitching in short relief than long relief.