Yoennis Cespedes has made quite a splash since leaving the communist paradise of Cuba, aligning himself with some very clever marketers and well-produced promotional videos. Scouts are raving about his power and his speed, and there is some anticipation of a bidding war for his services once he clears the regulatory hurdles that come with being a Cuban emigre.

As well there should be. While I was initially pessimistic about his past performances, and the projections that can be reasonably made from them, I discovered a few things about how the system works. Those influences were holding him back, so to speak, for what I’m pretty sure were bad reasons.  He looks like he should be a solidly above-average major league player – not likely to be a Hall of Famer and maybe not even an All-Star, but someone who could place in the top third of starting center fielders for the next several years.

I am not basing that conclusion on the videos prepared by his agents, nor on the reports of scouts. What I have, that no one else seems to have done yet, is to compile his performance history  from the Cuban Serie Nacional, along with the stats of everyone else in the Cuban leagues, and subject those statistics to the same kind of review we have for players in the American minor leagues. I have what I think is a complete record back to 2001 of Cuban batting and pitching performances, and fielding numbers from 2006 to now. You can see all of them by going through the ‘DTs by League’ tab; the drop-down menus will allow you to change the league to “Cuban Serie Nacional” (its at the bottom of the list) , to change the year (the Cuban League typically runs from November to April; the year given is the calendar year in which the league finished, so “2011” is the 2010-2011 season), to switch between hitting or pitching stats, and to switch between Real stats, Translated stats, and Peak Translated Stats. The latter isn’t as useful for Cuba as it is for American leagues, since there are so many players whose age is unknown  – at least, unknown to me. Even some of the ones we think we know turn out to be wrong. The drop-down menus also have a ‘Splits’ menu, but that won’t work with the Cuban stats – all I have are the top-line ‘ALL’ numbers.

Each of those pages are sortable. The stats for the entire league occupy the top of the page, with separate sortable stat boxes for each team down below. Each player is linked to his own page – accessible through the league pages, or through the search box at the top of any of the player pages. At least I think I have them all on their own page – the Cuban culture apparently takes a very lackadaisical attitude towards consistent spelling, which made it an enormous chore to link player stats from one season to the next together. I am almost certain that there is somebody whose stats are split in two because I didn’t catch on that Yoandry Malleta and Joandi Mayeta were, in fact, the same person. Having these pages gives us a chance to look at what Yoennis Cespedes has actually done on the baseball field in a competitive environment.

His real statistics are reasonably impressive – a consistent .300 eqa, averaging 35 HR per 650 PA (the sum lines are scaled down to 650 PA to ease interpretation), with excellent center field defense. He’s pulling 350-ish atbats a year, which – given that the Cuban season is 90 games, and the league leader in AB is around 380  – speaks well to his durability.

The 33 home runs he hit in 2011 represented a new Cuban league record. I have said before, that records, as often as not, are not the product of a great individual effort, but a good effort carried out in especially favorable circumstances. Cespedes’ home run record is undoubtedly the latter. Having the full stats in hand allows us to look at the trend in total HR hit over the last few years in Cuba: how they’ve gone from 669 in 2007, to 1192, then 1292, 1498, and finally 1449 in 2011, all with roughly the same number of games and plate appearances. From 2001 to 2007, no hitter in Cuba had more than 28 home runs; its happened nine teams in the last four years. In addition, Cespedes plays with Granma, which has had the highest park factors in Cuba over the last three years. (To be fair, he did hit 18 of his 33 home runs on the road, so the record is not simply a park effect). But a look at the leader board for home runs in 2011 makes it crystal clear:

Cespedes was tied for the record this year by Jose Abreu, who did it in 60% as mnay atbats, and there are a whole slew of players right behind them. This looks like a pretty normal leaderboard, not the leaderboard of a record-setting season – which is how you can be pretty sure the record really belongs to the conditions, not the individual. Give him credit for leading the league in HR, but leave the record talk out of it.

(Aside: you really, really should click on Abreu’s link and look at his numbers. If I were an MLB exec I’d be tempted to hire an extraction team to go in and kidnap the guy so he could play for me).

The thing is, while there are certainly some high quality players in the Cuban league, enough to fill out an All-Star team that is strong in world competitions, the quality depth just isn’t there – a problem that isn’t helped by the continual exodus of top players like Cespedes (let us please leave the moral issues, governmental ideologies, personal freedoms, what is right and what is wrong out of this; I am explicitly and only considering the baseball issue here). In past analyses I have graded the Cuban league, as a whole, to be on par with low A ball in the States. This means that the translation process is going to take some very big bites out of these Cuban statistics, which you can see for yourself in the Regular DT portion of his page and which I’ve reproduced here:

Yoennis Cespedes               Born 19851018 Age 25   Bats R   Throws R  Height 70  Weight 200   Regular DT
Year Team         Lge  AB  H   DB  TP  HR  BB  SO  R  RBI  SB  CS  Out  BA   OBP  SLG   EqA EqR POW SPD KRt WRt BIP Defense
2004 Granma______ CBA 300  73  17   4   8  23  86  40  34   3   1  233 .243 .302 .407  .246  35   5   0 -18  -3   4  77-DH   0
2005 Granma______ CBA 358  92  22   3  13  26  83  55  42   4   1  273 .257 .314 .444  .261  48  10   1  -8  -3   0  93-DH   0
2006 Granma______ CBA 360 101  23   2  19  31  68  66  58   6   1  262 .281 .344 .514  .289  59  17   3  -1  -1   0  87-CF  -3
2007 Granma______ CBA 361  91  22   2  17  30  84  67  57  12   5  278 .252 .317 .465  .267  51  15   8  -8  -2  -4  77-CF   8
2008 Granma______ CBA 378  82  14   1  18  21  79  57  52   3   2  301 .217 .258 .402  .225  36  13   3  -6  -6 -21  76-CF   9   1-LF   1
2009 Granma______ CBA 348  87  14   1  18  28  60  58  54   4   2  270 .250 .307 .451  .260  47  14   1   2  -2 -13  65-CF   2
2010 Granma______ CBA 358  93  19   3  14  30  70  61  45   4   1  267 .260 .321 .447  .264  48   9   2  -2  -2  -3  75-CF  -9
2011 Granma______ CBA 375  92  16   1  22  34  67  60  65   8   1  287 .245 .311 .469  .267  53  18   2   1  -1 -16  85-CF  16
                  --------------------------------------------------------------------------------------------------------
         Minors       591 148  31   4  27  46 124  97  85   9   3 2171 .251 .309 .451  .260  78  13   2  -5  -2  -7  97-CF   5

What we have is a guy who (over a 162-game season) has 25-30 HR power, which is worth roughly 15 runs more than an average major league player. The only evidence for good speed, which was on prominent display in his videos, are his fielding ratings – it is not apparent from his hit distribution or stolen base totals. Statistically speaking, he only rates as slightly speedier than average. His strikeout and walk rates both rate as “poor”, although his K rate has improved in recent years. Initially, at least, and especially so if a team tries to move him straight to the majors, the K rates are likely to be worse than this. The really bad mark on his record are his BIP numbers, which have become sharply worse recently, with a score at least 13 runs worse than major league average in three of the last four years. I am still in the process of understanding the way BIP fluctuates. The BIP score, by a wide margin, has the least continuity from one year to the next of the five component stats I have listed:  correlations are only about 0.4, instead of the 0.8 ratings the others enjoy. While there are tendencies for different hitters, there is a pretty good chance that this rating numbers will improve in the US.

So yes, the overall projection is for good power, low BA and OBA, and a good CF, with an overall EQA in the .260-.270 range.  Per the EqA report, the average EqA for major league center fielders last year was .269.

There is a major league player who is, statistically, quite comparable to Cespedes. He is also a center fielder, has a Gold Glove, and is only about 2.5 months older. Compare their 2011 DTs, enlarging Cespedes’ to the same plate appearance total:

.                    ab      h     db  tp hr  bb  so     sb  cs   ba     oba    slg     eqa

Cespedes       544  133  23  1  32  49   97   12   1    .244   .307  .467   .267

ML Player     562  162  25  2  25  31  100  11   5   .288   .325   .473  .274

The major league player gets more 24 hits, but gives most of that advantage back by drawing 20 fewer walks. He’s had major league EqAs of .246, .265, .262, and .274 in his career, and there is a near-constant expectation for him to break out and have a great season. Instead, we have a string of seasons which place him as the 7-10th best CF in the majors – very good, but short of All-Star caliber. The hype of what we expect from Adam Jones and the reality we’ve gotten seems to me like a very good lesson for Yoennis Cespedes.

When I run a projection for Cespedes – and, for that matter, Jones – I get a forecast that carries them from their current .270ish figure to something more .280ish. Combine that with being good-fielding center fielders, and you’re talking about 4-5 WARP, right on the border of All-Star status. A 4.3 WARP, which is his 50% projection, would have made him the 6th best CF in 2011, 5th in 2010, or 4th in 2009. The overwhelming majority of players with similar stats are in the majors; the Improve percentage and breakout/collapse ratios are both in his favor for the next few seasons.

Cespedes’ comps are interesting, in that there isn’t a lot range to it. None of them are in the Hall of Fame, but a couple of them get mentioned in Hall discussions. Seven of the ten were very, very good – and the remaining three did just about nothing, with no gradation between the two groups.  His ten best comps (based on performances from ages 23-25) :

“Year” corresponds to the “current season”; WARP and VORP are career totals; EQAlast3 is the players translated EQA for the three seasons prior to “Year” (e.g., 1973-75 for Hendrick), while EQAnext3 is for Year plus the next 2 seasons (1976-78 for Hendrick); the component scores are also for the prior three seasons.

George Hendrick, not surprisingly, also comes up as Adam Jones’ #1 comp.

One thing that is always a question with Cuban players is age – how much do these stats change if Cespedes is not actually going on 26, but is in fact several years older? It actually isn’t that bad for him, compared to other players who were reportedly something like 22 when coming to the States. He’s reached what you might call the plateau portion of the aging curve, where expected performance stays fairly level for about six years. Changing his current age had almost no affect on the projections for 2012 or 2013; where it does have an affect is in how long he can play before he goes into the downhill portion of the aging curve. You can see in his projection that the breakout/collapse ratio goes below 1 at age 29, and continues getting worse; that’s a good indication of where his comps have tended to lose it.

Looking back at trades and moves since Christmas:

Red Sox: Traded Josh Reddick, Miles Head, and Raul Alcantara to Oakland for Andrew Bailey and Ryan Sweeney. Signed Rich Hill.

By adding Bailey to their bullpen, who should immediately take over as the closer (sorry, Mark Melancon), I’m reasonably certain that at least one of Alfredo Aceves or Daniel Bard will join the rotation. Its been talked about all off-season, but in my experience these discussions happen about five times as often as the move actually happens. Good relievers turning to starters have a pretty good track record – Ogando last year, Ryan Dempster, CJ Wilson – and I would expect Bard, Aceves, or both to do fine. In right field, I suppose Sweeney just slots in directly for the playing time I expected would go to Reddick. In the mean forecast, I actually have Sweeney being a touch better than Reddick for 2012 – but that is conditioned by Sweeney having some demonstrated fragility, a much lower chance of surprising you to the upside this year, and a lower chance of improving in future years. Sweeney adds to the collection of OK but unexciting outfielders the Red Sox have in-system: Ryan Kalish (who will miss the first two months or so), Dan Nava, Alex Hassan.

Hill is a TJ-reclamation project five years removed from a good season, and I don’t expect him to contribute.

Cubs: Signed Andy Sonnanstine.

In the Tampa Bay system, Sonnanstine was the beneficiary of solid defensive teams that made him look, at times, almost respectable. Wrigley Field is not going to be his friend.

White Sox: Traded Carlos Quentin to San Diego for Simon Castro and Pedro Hernandez. Traded Jason Frasor to Toronto for Miles Jaye and Daniel Webb.

I am not sure what the Sox are trying to do with this, unless they really think that Quentin is damaged beyond repair or such head case they no longer want to deal with them. Who they expect to play the outfield for them has me baffled – I suppose they are committed to Alejandro De Aza and Dayan Viciedo, but Brent Lillibridge (or Adam Dunn(!)) are the only other players in-system with outfield experience and a projected EqA north of .245, which equates to “remarkably thin”. I had them as the best contender to the Tigers for the AL Central, but they’ll have to shore up that outfield. I’m wondering, given their experience with Viciedo and Alexei Ramirez, if they aren’t making a strong for Yoennis Cespedes.

Of the pitchers they got, I’d expect Castro (who reached AAA wit the Padres) and maybe Hernandez (who made AA) to make appearances in the majors this year, but not to remain up for very long. Jaye and Webb are much farther away, and none of the four have pitched like a sure-fire major leaguer.

Yankees: Signed Andruw Jones.

Jones steps in for what figures to be another 200 or so PA, backing up the heretofore durable Yankee outfielders and occasionally stepping in at DH. He’s a step up from Justin Maxwell, who pretty much drops off the depth chart.

Athletics: Traded Andrew Bailey and Ryan Sweeney to Boston for Josh Reddick, Miles Head, and Raul Alcantara. Signed Jorge Soler.

Reddick is the only player here i expect to see in Oakland this year. He’s a pretty average outfielder, roughly equal to Sweeney for this year, but a fair bit more growth potential in the power department. Head turned in a really nice half-season in the Sally League last year, but none of his other stops point to a major league first baseman. Alcantara was in the Gulf Coast League last year, and did well enough for me to consider him a decent long-term prospect – but it’s a huge gap between the GCL and the majors.

Padres: Traded Simon Castro and Pedro Hernandez to the White Sox for Carlos Quentin.

I had Castro slotted for about 10 starts this year, which I’ll now pass on to Joe Wieland (who I like better, anyway) instead. I’m assuming that Quentin locks down left field, which should kick off a big fight between Yonder Alonso, Kyle Blanks, Jesus Guzman, and Anthony Rizzo for first base, a fight that could, in a stretch, also draw in Will Venable or Chris Denorfia. That’s a fight that Alonso probably wins, although the projections for all of them are close enough that the hottest March hand could take it. Quentin gives the Pads a proven power bat, but a) they look like a distant fifth-place team, b) he’s brittle, and c) has a pretty lousy projection going forward for a 28-year-old (probably because of injuries and lack of speed) …I don’t see how this fits in as part of a plan.

Giants: Signed Boof Bonser.

Hasn’t had a positive RAR since 2007.

Blue Jays: Traded Miles Jaye and Daniel Webb to the White Sox for Jason Frasor. Signed Darren Oliver. Signed Aaron Laffey.

When I first filled out a depth chart for the Blue Jays, I had real problems with the Toronto bullpen, and I wound up assigning 20 innings apiece to seven different guys with lousy (4.50+ ERA) projections. Frasor and Oliver – not so much, Laffey – wipe out most of those and cut a quarter run off the projected bullpen ERA.

Nationals: Signed Michael Ballard.

10 years, 254 Megabucks.

Ten years.

That’s an eternity in baseball, particularly when you are dealing with a player who has already passed the age of 30.

I have a new card for him up here. Lets start with the projections for this year and beyond, down at the bottom of the page:

Year Age PA  AB   R   H   DB TP  HR RBI  BB  SO SB CS  BAvg OBP  SLG  EqA  RAR  WARP  Defense   MJ Brk Imp Col Att Drp
2012 32  596 520  90 160  25  1  34  98  71  58  9  4 .308 .396 .556 .329  57.3  7.1 139-1B  3 100   9  38  18   3   1 
2013 33  592 518  88 159  25  1  32  96  70  58  7  4 .307 .394 .544 .326  54.1  6.7 138-1B  3 100   2  32  25   9   1 
2014 34  575 503  84 153  24  1  30  90  68  59  6  3 .304 .391 .535 .323  50.7  6.3 134-1B  3 100   3  25  31  15   6 
2015 35  553 483  78 144  23  1  28  84  66  56  6  3 .298 .387 .524 .319  45.9  5.8 129-1B  3 100   6  19  42  25  13 
2016 36  485 423  69 126  20  1  25  75  58  49  5  3 .298 .388 .527 .319  40.6  5.0 113-1B  2  99   3  15  52  37  24 
2017 37  447 392  64 116  18  1  23  69  52  46  4  2 .296 .383 .523 .316  36.1  4.5 104-1B  2  99   0  12  58  45  33
2018 38  407 357  56 104  17  1  21  62  47  42  4  2 .291 .378 .521 .315  31.8  4.0  95-1B  2  97   1   8  63  51  42
2019 39  360 316  50  91  15  1  18  54  41  37  3  2 .288 .375 .513 .311  26.6  3.4  84-1B  2  98   1   6  69  61  53 
2020 40  300 263  39  76  10  0  15  44  35  32  2  1 .289 .377 .498 .309  21.3  2.7  70-1B  1  96   0   4  81  71  63
2021 41  248 221  31  63   9  0  12  36  26  27  1  1 .285 .363 .489 .301  15.2  1.9  58-1B  1  96   0   2  88  81  71
.

(I’ve cut some columns to make it fit here, and added more projection years than the pages are going to have). The top-line projection is for him to steadily drop from an EqA of .330-ish with a 7 WARP down to .300 with a 2 WARP – my gut feeling is that the long-term projections underestimate the EqA decline while exaggerating the PA decline, but that the overall level comes out about right. There’s 47.4 WARP showing on the board for the duration of the contract. The marginal value of those WARP needs to be about $5.35 million over the life of the contract for it to break even, which is pretty close to the current rate. If you allow for a 5% growth in the marginal value per year, then you only need to start around $4.15M to break even; since current marginal values are probably around 5.25, there’s a little (very little) bit of room to come up short in the expectations. The value looks OK for the expected production, as a total. Even for that, though, you are looking at a strong, strong likelihood of overpaying for the years at the end of the contract…you just have to hope you come out far enough ahead in the first few to make up for it.

But the risks, oh, those risks. The drop rate (“Drp”), which is the percentage of his comparables who are completely out of the league, is already up to one-third through six years; that number is a big part of why the expected plate appearances drop so steadily through the forecast. His breakout (Brk) scores, the chances of turning in a performance well above his 2009-2011 baseline, is only in the single digits throughout the contract; his collapse rates, the chance of underperforming that baseline, grows steadily, as expected. There is far more downside risk than there is upside, and the contract doesn’t leave much downside room to keep from becoming a stinker.

There is also the case – already factored in to the total forecast, but something I’d like to look at a little closer – that we may have already seen some decline. Here’s a cut from the regular translation portion of his card:

Albert Pujols                  Born 19800116 Age 31   Bats R   Throws R  Height 75  Weight 230   Regular DT 
Year Team         Lge  AB  H   DB  TP  HR  BB  SO  R  RBI  SB  CS  Out  BA   OBP  SLG   EqA EqR POW SPD KRt WRt BIP 
2009 St_Louis____ NL  571 191  36   1  50 109  57 129 144  16   5  395 .335 .442 .664  .357 150  31   0  16  12   1 
2010 St_Louis____ NL  585 189  32   1  46 103  66 122 128  14   5  409 .323 .423 .617  .340 138  26  -1  14  10   1 
2011 St_Louis____ NL  573 182  24   1  42  64  50 119 114   9   1  402 .318 .384 .583  .320 117  22   0  17   2  -3

.

The last five columns I’ve left in, after EqA and EqR, are component breakdowns. They are rate statistics, not totals, measured in runs per 650 plate appearances. To digress a moment, there was a comic book character in the DC universe by the name of Ultra Boy, who had all the powers of Superman – but he could only use them one at a time. If he turned on the strength, he wasn’t invulnerable, if he was invulnerable he couldn’t use the heat vision, if he was using X-ray vision he couldn’t fly to get a good vantage point, and so on and so on. I think Peter Petrelli had the same thing in Heroes. That’s kind of the idea here, with Pujols standing in for Superman. I have taken a perfectly average player, and replaced his averageness with one aspect of Pujols at a time. The POW column gives Mean Joe Mean Pujols’ power – mostly his home runs, with some doubles counting in as well – while leaving everything else perfectly average. SPD does the same thing for speed – stolen bases, triples, some doubles. KRt is strikeout rate, WRt is walk rate (which, for this purpose. includes HBP), and BIP is essentially singles.

What I really want to focus on is the drop in POW – from +31 in 2009, to +26 in 2010, and to +22 last year, to find players who underwent similar changes at the same ages (29-31), and to see what they did in the future. I’ve run cards like this for every player since 1954 (which you can find, for now, by going through the search bar at the top of each player’s page), so I started looking for players who had POW scores at ages 29-31 that matched Pujols’ scores, plus or minus some value. There weren’t any players found when that plus or minus value was 0, 1, or 2, which isn’t surprising since a +20 power score is a pretty high bar. I started to get some nibbles at +/- 3, a few more at 4, and by +/-5 I had twelve players to work with. Chronologically, they are

Ted Kluszewski  1954-56    34, 24, 18
Stan Lopata, 1955-57       32, 27, 19
Bill Skowron, 1960-62      26, 23, 23
Lee May, 1972-74           28, 27, 21
Johnny Bench, 1977-79      31, 29, 17
Dwight Evans, 1981-83      27, 23, 17
Steve Balboni, 1986-88     28, 22, 24
Barry Bonds, 1994-96       32, 22, 27
Jeff Bagwell, 1997-99      31, 21, 27
Mo Vaughn, 1997-99         28, 28, 22
Manny Ramirez, 2001-03     37, 27, 21
Troy Glaus, 2006-08        26, 21, 20
ALBERT PUJOLS, 2009-11     31, 26, 22
.

(Note that for the older players, pre-1980, the “new” design player cards haven’t caught up yet.)

These 12 players, between them, were able to meet or beat their worst age 29-31 POW 28 times over the remainder of their careers, an average of 2 and a third times per player, which doesn’t sound too bad. Nine of those 28, however, belong to Barry Bonds; six more belong to Manny Ramirez, which means that more than half of the good results come from two players with PED issues. The remaining ten players tally only 13 seasons, just a little over 1 per player

Kluszewski had a 21 in half-time play for the Angels in 1961
Lopata was done.
Skowron never reached 20 again.
May had a 29 in 1976, age 33, and had a couple more high-teens.
Bench bounced back with a 30 in 1980 and a 19 in 1981, so he gets credit for two.
Evans went 21-14-18-22 over the next four years, giving him three hits.
Balboni put up a 28 and 36 in the next two years, and then was out of the league.
Bonds went ape on the league for the next decade, including an incredible 68.
Bagwell went 26-21-18-20-12 for the remaining five full years he played, that counts for 2.
Vaughn had a 22 in 2000, missed 2001, had another 22 in 2002, then an 8 and then was done, so that counts for 2.
Ramirez had a 31-34-33 over the next three years, and then a 21, 25, and 33 later on, for a total of 6.
Glaus was pretty much done.

This is, to some extent, factored into the general projection, although the comps there are based on similarity across all performance sectors, not just power (as well as height, weight, position). Pujols has significant advantages over many of the players on this list in terms of speed (just by being average, instead of -5 or -10), strikeout rates (very strong +15 typically, compared to -22 from Balboni), and walk rates (although the drop from 12 to 10 to 2 is troubling), which should help him remain a star player even if his power drops into the mid-teens. Compare him to someone like Balboni, for whom power was the only component worth anything; and even at a +36 the major leagues didn’t think it offset his high strikeouts (-29), lack of speed (-8), and lack of singles (-23) enough to justify bringing him back for more. We would expect that kind of broad toolset to age better than a narrow one.

A quick look at how the last 12 years of pennant races would have looked like with a second wild card:

In six of the last eight NL races, the carte deux team finished just 1 game behind the actual wild card winner. From 2001 until now,  the fifth-best NL record never been more than 4 games behind the wild card.

By contrast, this year’s race, between the Rays and Red Sox, is the only time since 2000 that the fifth-place AL team has finished within 1 game of the wild card. In eight of the last 12 years, the new card winner (assuming things played out as before) would have trailed the real winner by 5 or more games, with the 2001 Twins finishing a whopping 17 games behind the “102 win but still just a wild card” A’s.

Given that knowledge, you’d think the fifth-best NL teams would have a better record than their AL counterparts, but you’d be wrong. The AL wild cards averaged 95.42 wins to the NL’s 91.25 from 2000-2012, and their “tame cards” still led by 89.84-88.67.

In 2007 and 2002, the “Drive for Five” (“Drift for Fifth”?) would have ended in a tie in the AL, between Detroit and Seattle in ’07 and between Boston and Seattle in ’02.

In the 2007 NL, the Padres and Rockies finished the regular season in an 89-73 tie for the wild card, forcing a one-game playoff between them – exactly what would have happened under the new system.

In seven of the 24 leagues between 2000-2011, the second wild card actually had the fourth-best record in the league, beating out at least one of the division winners. In two more seasons the WC2 was tied with one of the division winners – so that quite often the “fifth” team has as good a claim to the playoffs as one of the division winners.

In the 2011 AL, 2010 AL, 2009 NL, 2008 AL, 2007 NL, 2006 AL, and 2002 NL, the two wild cards were from the same division. Getting the wild cards coming off a one game playoff, with likely disruptions to their starting rotation and bullpen, just got a little more important, relative to facing a rested division winner. Will the team with the best record still get bumped from the wild card if they come from the same division?

Similarly, holding the tiebreaker for a division lead in a “one team wins the division, one gets the wild card” scenario just got more important.

The Giants have been the “last team looking in” three times since 2000, in 2009, 2004, and 2001. The Red Sox held the distinction in both 2010 and 2011, plus they were tied for it in 2002. The Red Sox would also have had to defend their wild card win five times, more than anyone else. The Padres, Phillies, Dodgers, and Indians have been in the situation twice. The Mariners were that team once, and tied for being that team two more times.

NL
Year WC WC-2 GB
2011 STL (90) ATL 1
2010 ATL (91) SD 1
2009 COL (92) SF 4
2008 MIL (90) NYM 1
2007 COL (89+) SD 0 (not counting playoff)
2006 LA (88) PHI 3
2005 HOU (89) PHI 1
2004 HOU (92) SF 1
2003 FLA (91) HOU 4
2002 SF (95) LA 3
2001 STL (93) SF 3
2000 NYM (94) LA 8

AL
2011 TB (91) BOS 1
2010 NYY (95) BOS 6
2009 BOS (95) TEX 8
2008 BOS (95) NYY 6
2007 NYY (94) Det/Sea 6
2006 DET (95) CWS 5
2005 BOS (95) CLE 2
2004 BOS (98) OAK 7
2003 BOS (95) SEA 2
2002 ANA (99) Bos/Sea 6
2001 OAK (102) MIN 17
2000 SEA (91) CLE 1

Not that it shows on the site, but its been a pretty busy week for me with baseball ideas.

The first priority has been to clean up and expand the DT pages (“By League” on the subject line). I’ve gotten a year key added in, and it does seem to be working for the years that have split data – from 2005 to 2011. I am getting things  set up to have access to the full season DTs for all leagues going back to the Seventies.

Those files will include at least some winter leagues. The Arizona Fall League just wrapped up, and I’ve got their data entered in – it should come through on the league pages shortly. The Central and Pacific leagues in Japan wrapped up their regular seasons near the end of October, and I’ve gotten all of their data entered – that task certainly is much easier now that NPB has gotten their English-language site back up. And I’ve downloaded Cuban data for the last few years, but haven’t been able to process it yet – something I’m eager to do with recent attention on Yoennis Cespedes.

I’ve also been going through my programs and cleaning them up. Some of this code is going on 25 years old now, and it is full of blocks that aren’t commented as to what its doing – and my ability to just remember things like that isn’t what it was back then. There are blocks that don’t do anything any more but are still there, blocks that do something only to have it redone a different way immediately afterwards, meaningless variable names, things that are just plain sloppy. I know that I did it that way back when because a) I didn’t necessarily know any better, and b) I was usually in a big hurry to move on to the next thing. I look at some of this code now, especially in relation to the standards I have at work with NOAA, and its kind of embarassing, even if I am the only one who sees it.

One big clean-up was the way the program process “peak” translations. In a normal translation, you are adjusting a (usually minor-league) performance into an estimated stat line for the major leagues. A peak translation is similar, except that you are building an estimated stat line for the major leagues of some future year, when the player is at his peak – an estimate of how good he can be, rather than how good he is now. The primary determinant of that is still current ability level plus age, but the articles that Rany posted recently at BP, some looks I was taking at Bryce Harper, and stats I ran to validate league difficulty levels all go me thinking of other approaches to the problem. One big change is to yank all the future code out towards the end of the other program, letting the regular DT play all the way out before trying to adjust it; I had always jumped in early on, and sometimes I wound up with conflicts between the ‘normal’ and ‘peak’ DT levels. I also changed the adjustments from a system that was primarily multiplicative – “power increases 20%” – to one that is primarily additive – “power improves by 45 points”. That enormously simplified problems I had with overestimating  players who had monster minor league seasons. Those changes should be apparent in the stats pages very soon.

I’ve also been validating projection systems from the 2011 season. While I’m pleased with how my system (which ran with some of Nate Silver’s ideas on PECOTA, threw out some of them, replaced them with some of my own tools and approaches, resulting in a chimeric Sildavenverport monster) graded, and I was also pretty shocked at just how little difference even the most complex systems made when compared to an ordinary three-year average.

One of the biggest plays in game two came in the ninth inning, when Josh Hamilton lofted a fly ball to right field. Ian Kinsler, on third base, at the time, scored; Elvis Andrus, on second, tagged and went to third. He’d score the winning run on another sacrifice fly from the next batter.

We can construct a table of values to show us how many runs scored, on average, from every base/out situation. These are the numbers for the major leagues in 2011:

            0        1       2      1-2      3      1-3     2-3    1-2-3
0 outs     0.480   0.849   1.062   1.431   1.311   1.680   1.893   2.262
1 out      0.258   0.502   0.648   0.892   0.898   1.142   1.288   1.532
2 outs     0.097   0.218   0.314   0.435   0.354   0.476   0.572   0.693

Hamilton’s fly took the Rangers from a situation that averaged 1.893 runs, to one that had 1 run definitely in, plus an average of 0.898 runs more to come. That’s a gain of just 0.005 runs, which makes the play seem almost inconsequential. It goes to show that flying out to get the run has a lot less intrinsic value than banging out a hit would have carried.

That’s one way of looking at it. Compared to just striking out, though, or popping up or doing anything that would have left the runners in place, the real result was a clear gain of 0.61 runs. A quarter run of that comes from Andrus’ getting from second to third, which perhaps should be credited more to him than to Hamilton, but that is besides the point.

In turns out that a play like this should come as no surprise. Perhaps because he did frequently have players like Elvis Andrus on base in fron of him, Josh Hamilton was the #1 hitter in the majors in terms of value gained from his outs, and by a wide margin:

Josh Hamilton, Tex     20.1
Derek Jeter, NYY       16.2
Juan Pierre, CWS       15.7
Omar Infante, Fla       15.5
Carlos Lee, Hou          15.2
Hideki Matsui, Oak     14.9
Angel Pagan, NYM    14.8
Ichiro Suzuki, Sea      14.6
Elvis Andrus, Tex       14.5
Alcides Escobar, KC  14.4

To get to these figures, I used the run matrix above to define a baseline value of a simple out – one with no runner advancement. That was compared to the real value after the play was over whenever the batter hit a sacrifice fly, a sacrifice hit, or had a hitless atbat. Reaching on an error counts in your favor – looks like an out in the boxscore, plays like a single or better as far as the game is concerned. Grounding into a double play sticks a minus on you – from the game point of view, and Earl Weaver’s, you should have just struck out. So does hitting a fly ball deep enough to tempt a runner into trying to advance, but not enough for him to succeed.

There is no question that these values are dependent on the men on base for you; I’m only trying to reconcile what happened while hitters were up, not what they might have done with “average” baserunners. And there is also no doubt that my no-advance baseline can be questioned.

As you can see from the above list, it helps on this measure to be fast – that’s going to help you reach on errors and avoid double plays, and those will help you score well by this metric. It clearly isn’t the only thing, though, since that adjective does not remotely apply to Lee or Matsui. A look at the bottom ten, though, does reinforce the speed angle:

David Ortiz, Bos            -5.1
John Buck, Fla               -3.3
Ryan Adams, Bal          -3.0
Matt Holliday, StL         -2.8
Matt Wieters, Bal          -2.0
Pedro Alvarez, Pit         -1.9
Brian McCann, Atl        -1.9
Brad Hawpe, SD           -1.9
Matt Domiguez, Fla      -1.5
Brian Bogusevic, Hou   -1.3

An big, old DH leads the way, three catchers, an out-of-shape third baseman, three short-time players unlucky enough pull some big negative plays.

The World Series also offers quite the clash of teams, in this stat. The Rangers are the only team to have two players in the top 10, with Hamilton and Andrus, and were also the only team in the majors to have no negative players at all. Its not too surprising, then, that they accumulated a major-league best 103 runs above the default, beating out the Mets by two and a half runs.

At the other end of the spectrum, Holliday’s place in the bottom ten is a clue to the Cardinals, who finished with a major-league worst total of just 46 runs above baseline. Holliday (-2.8), David Freese (-0.7), Colby Rasmus (-0.2), and Albert Pujols (-.005) gave the Cards four regulars who just weren’t going to advance anybody without a hit. Their best player, semi-regular David Descalso, had just 8.5 runs, equaling the worst mark for any team leader.

The team totals look like this:

TEX         103.4
NYM       100.9
CIN            97.5
TOR          96.4
KC             95.0
COL          94.2
SEA          93.6
SF             89.8
WAS         88.7
OAK         86.3
MIL          85.1
PHI           82.9
HOU         82.8
LAD          82.1
TB            81.1
FLA          80.7
DET         79.9
SD            79.4
PIT           78.8
CHC         75.6
ATL         74.9
CWS         74.2
NYY         73.7
BOS          71.2
MIN          71.2
ARI          62.6
LAA         62.5
CLE         61.2
BAL         50.0
STL         45.6

Do the team’s abilities in making productive outs have any relationship with their projected run totals? The question is slightly confused, in that SH and SF are included in EqR, but there was essentially no correlation between the difference in productive out deltas and EqR deltas – and what little correlation there was was slightly negative. Boston, the Yankees, and St. Louis had top-five offenses, despite getting low advancement value on their outs, because of the frequency and strength of their non-outs. If you get enough hits and homers, it doesn’t matter how often someone moves from second to third on a grounder.

The other night, I finally got myself together to hunt through the grimmeries and other web pages of lore to find the right set of incantations that would allow me to produce a graphic more like what I’ve always wanted for the Postseason Odds Reports;. The new versions look something like this

Urg. Go here to see what it is like when it isn’t squished by the frames.

So what’s different from the old ones? A clear label for the various months (the vertical bars are the first of each month) is a really big deal, to me. The size is fixed, and shows you where the end of the season is; the old version always ended with “today”, and gave no indication of how much time was left. The flat-lining teams aren’t lost in the borders of the frame anymore, although that’s no big deal to me – still, one person I showed an early version to liked it better that way.

This improvement comes pretty late in the season, though, well after the excitement has gone out of reading these charts. Three of six division have a team who’ve reach ed a 99% chance of making the playoffs; two more are in the 88-89% range; the wild cards are also right around 99%. The one I’m showing here is the only race that, by my standard method, still has a strong pulse.

For those new to the site, the postseason projections are the result of a simulation I run every day, where I play out the rest of the season a million times. The team’s current record is known, and is always the starting point for the study. There are three different schemes that I use for rating the quality of teams, which is used to set the likelihood of winning each game, methods which have different strengths and weaknesses. All of the graphs are available by clicking on the tagged division name on each page.

The standard version uses the statistics of the team during the season. At the start of the year in this scheme, every team within a division has an equal chance of making the playoffs, chances which are only changed by the actual performance of the team. The most important factor is the team’s actual record. We also consider what a team’s record should be based on the runs they have actually scored and allowed, and also what the record would be with the expected number of runs and runs allowed, based on their batting and pitching statistics. The quality of opposition – opponent’s hitting and pitching – also figures in. I also incorporate some terms to regress the team towards the mean, and a random variable meant to simulate the fact that my rating the Red Sox as a “.585” team is just  an estimate; they may really be a .650 team or a .500 team. The advantages of this method are that it makes the fewest assumptions about a team, doesn’t pre-judge, and can be readily applied to past seasons. The main disadvantage is that it does not adapt to changes in team personnel (such as injuries or trades); a team who holds a fire sale can be completely different from the team who accumulated the statistics used to rate the current version.

The second version, the “forecast” version, uses the forecast statistics of all of the players on the team to estimate the team’s quality. Unlike the standard version, this one can react immediately to changes in personnel. The biggest  downside is that it takes an awful lot of work and time to maintain a full set of forecasts and depth charts and keep both of them current throughout then season. I managed to do it last year; I haven’t come close this year. You are also not going to be to do any better than your chosen forecast method, and you are going to introduce human bias with the playing time estimates. I would also add that it is nearly impossible to make an honest depth chart for a prior season, so running it for a different year is futile.

A third version was mostly dreamt up by Nate Silver,, who worked out a scheme for using Elo ratings; for baseball. Leaving aside the conversion of winning percentage to an Elo score, this version is similar to what you would get if you ran the standard version as some kind of weighted average, treating recent results (wins, runs, etc) as more valuable than more distantly past results. It even uses prior season data to start the teams with unequal chances, and it does do a better job than the standard version of reacting to changes in the team.

Generally speaking, I do prefer the original version, despite its flaws.

Oh, and yes, earthquake yesterday. Had I been elsewhere, I might have considered it exciting. Being on the fifth floor of an eight-story building, with who knows what standards of earthquake resistance – that leaves plenty of room to fall and still have a lot left over fall on top of you. I headed for the stairs with my phone in my hand, but my glasses on the desk; I realized I didn’t have them about halfway to the stairs, but since that was probably during the peak of the shaking I wasn’t inclined to turn around (especially since there were more people behind me moving forward). It was funny that I could get internet service but not phone service, and from the parking lot I pulled up the USGS earthquake site. I could not believe we had a sixish quake around here. Everything turned out fine, though, and the house didn’t have anything worse than my wife’s earring/jewelry box toppling over. Even the cats didn’t seem too upset.

The question I have in the debt limit crisis – which, removing the passionate politics from the equation, you have to admit was an incredible large-scale psychological study, sure to spur armchair analysis for years to come – is the extent to which the president campaigned for a cause other than compromise.

From my, admittedly left-of-center vantage, the compromise was more important than the subject matter at hand. I had previously characterized Obama as the character you see in many war/sports movies – the one who tells the troops not to fire too soon, to wait for it, wait for it, waaaiiittt ffffoooorrrr iiiittttt….NOW!

Now I find myself with a different war move characterization…that of Colonel Nicholson, the commanding British officer played by Alec Guinness in “Bridge on the River Kwai“. Nicholson so set his sights on a goal – building a proper bridge  – that he became completely divorced from the larger conflict, arguably becoming far too accommodating to his enemy. I can’t say that I know how that possibility – subsuming yourself to a different goal, losing yourself in a different reality from everyone else around you –  will eventually lead. Neither could the authors, as the ending for Nicholson was quite different in Boulle’s book than it was in Lean’s movie.

But overall, my impression of events matches those of Major Clipton, who had the last word in Kwai:

Madness. Madness!

On June 14th, the standings in the NL LABR looked like this:

1 Baseball HQ (Doug Dennis)                   97.5
2 Baseball Info Solutions (Steve Moyer)   95.5
3 Rotoworld (Wolf/ Colton)                        81.0
4 Hardball Times (Derek Carty)                80.5
5 ESPN.com (Tristan H. Cockcroft)          75.5
6 RotoWire (Dalton Del Don)                    74.5
7 ESPN (Eric Karabell)                              67.0
8 Sandlot Shrink (Bob Radomski)              64.5
9 Razzball (Rudy Gamble)                         62.5
10 NFBC (Ambrosius/ Childs)                   58.5
11 Yahoo (Brandon Funston)                     56.0
12 USA TODAY (Steve Gardner)            49.0
13 Baseball Prospectus (Clay Davenport) 48.0

This was a deeply embarrassing place to be, not the least because I was the league’s defending champion. I had also been very high on my team coming out of the draft, projecting myself as having a first-place team. Now it can certainly be argued that many players weren’t as good as I thought they were, but I truly believe that the main reason for this standing was injuries. The injury bug hit my team early and often; at one point in late April, over more than $120 of my $260 worth of players was off of major league rosters, either on the DL or in the minors – a list that went something like Ryan Zimmerman, Zach Greinke, Angel Pagan, Ubaldo Jimenez, Andrew Cashner, Barry Zito, Roger Bernadina, Jeff Keppinger, Ross Ohlendorf, and probably some others I’ve (no doubt chosen to) forget about.

Most of those players are back, and while they’ve been slow to regain their full form the change in my fortunes over the last six weeks has been most gratifying:

1 Baseball HQ (Doug Dennis)                     102.0
2 Baseball Info Solutions (Steve Moyer)      86.0
3 Rotoworld (Wolf/ Colton)                           78.0
4 Hardball Times (Derek Carty)                   77.0
5 Baseball Prospectus (Clay Davenport)  70.5
6 ESPN.com (Tristan H. Cockcroft)             67.5
7 Sandlot Shrink (Bob Radomski)                 66.0
8 ESPN (Eric Karabell)                                 65.5
9 NFBC (Ambrosius/ Childs)                        64.5
10 Razzball (Rudy Gamble)                          61.0
11 RotoWire (Dalton Del Don)                      60.0
12 USA TODAY (Steve Gardner)                58.5
13 Yahoo (Brandon Funston)                         53.5

I could conceivably make up another 10 points or so, giving me a realistic shot at third place, which would be a very satisfying performance – a very respectable title defense. I could, just as easily or even easier, find myself working back down through the pack. Being able to turn this one around elevates what was already a strong fantasy year for me:

LABR, currently 5th/13

MABL (work league, both leagues, fixed price menu of players) – currently 1st (of 12)

Groundhog League (NL E and C only, lots of odd rules) – currently 1st (of 9)

Murder City (AL) – currently 3rd/12; not likely to catch both teams ahead of me, but a pretty good lead on the 4th place team.

Murder City Lite (NL) – currently 2nd/14, good chance of winning it, as the lead has bounced back and forth

The best case scenario from here looks like 3 wins, a 2nd, and a 3rd, which would be easily my best overall year for drafting. The last three mentioned leagues are money leagues, so there’s a good chance that October will show a nice profit. You could say that I am very, very pleased with how my projection routines ran this year.

So the Mets have traded off Carlos Beltran, and their reward is a 21-year-old pitcher in high-A by the name of Zach Wheeler.

Wheeler comes from the Giants, and is very highly rated as a prospect…at least by the scouts. I’m here to judge him solely by the numbers. And I will say, in preface, that numbers are somewhat more likely to lie with a minor league pitcher than they are with a hitter. A hitter, minor league or otherwise, cannot dictate the action; he has to take whatever is thrown to him, and if he can’t hit a slider he can expect a steady diet of them, and the farther up the ladder he goes the truer that gets. A pitcher, by contrast, does have control of what he’s doing; he can, and does, work on secondary pitches, throwing pitches in minor league situations that he would never consider in the majors. This can have an extremely distorting effect.

The DTs don’t particularly like him. His strikeout rate is good, but hardly great; while he has struck out an impressive-looking 10 batters per 9 innings in San Jose, the league as a whole whiffs 8.33 times. His walk rate, 4.8 per 9, is pretty bad, almost disqualifying as a prospect, yet represents an improvement over the prior season. One thing that I think goes in his favor is that he has a huge platoon split; he gets absolutely clobbered by lefties, but has been right dominant against the other side. At a minimum, it provides a major league role that he could fill, even if he never does learn how to get the sinister ones out.

One type of translation that I produce, which I don’t have posted anywhere on-site, is what I call a “no-dif” translation. A nodif translation only tries to adjust for the offensive playing environment – the league offense and the home park, but not the difficulty level. That means, in nodif, that a perfectly league average pitcher will always translate to a 4.50 ERA, and 9.0 hits per 9 innings, 1 home run, 3 walks, and 6 strikeouts – in the AL, the NL, the Pioneer, Japan, in Denver or San Diego. The usual use is for making comparisons between leagues to find out just what the difference between AAA and the majors is, but I’ll use it here to look for similar pitchers.

My database on players goes back to 1978. When I look for 21-year-olds pitching in high-A, I get 2649 players. I’ll mandate 75 innings (Wheeler has 91) – that cuts it to 991. Wheeler’s a starter, so let’s put in a 12 start minimum, and that gets us to 835. Wheeler’s no-dif K rate comes out to 7.6 per 9 innings; we’ll select anyone between 7.1 and 8.1, and that leaves us with 113. Wheeler’s adjusted walk rate is 4.1. We’ll look at anyone with an adjusted walk rate of 3.5 or worse, and find that we have 40 pitchers left. Specify right-handed, and you’re down to 27.

Of those 27 pitchers, 16 never pitched in the majors. Four of them are currently active in the minors – Cody Scarpetta, Steven Johnson, Bruce Pugh, and Clint Everts – and could conceivably add to the list.

Of the 11 who reached the majors, five had minimal time. Chris Bushing pitched four innings over six games, Chuck Malone had seven innings, Steve Watkins and Kerry Woodson (not Wood) had 14 apiece, and Marc Kroon pulls in with 27. Woodson is the only one of the bunch who pitched well enough that you wonder why he didn’t get more of a chance. The five of them, combined, pitched in 58 games, starting just 1, and earned a total of -0.8 WARP3.

The list includes Erik Hiljus, who logged 124 career innings and a 0.3 WARP. Next up is Billy Buckner, with 138 innings and a -0.8 WARP. And there’s Wayne Gomes, who relieved 321 times in the majors, to whom I give a -2.3 career WARP3.

Eight of the 11 major leaguers, then combine for -3.6 WARP spread over 449 games and 696 innings. That leaves three pitchers we’ll call success stories.

Success #3 may be a highly premature call, since he’s only pitched 23 innings in the majors. But with a positive 1.3 WARP3, he’s already in third place on this list. Javy Guerra stuck his foot in the Dodger’s revolving door at closer and, for the moment, has brought it to a stop.

Success #2 is Roger Pavlik. Pavlik only pitched three full seasons in the majors, since he had enough injuries to earn loyal customer discounts with his local orthopedic surgeon. He did have two season with WARP above 4.0, and went 15-8 one year despite an ERA of 5.19 ( a mark which was slightly better than league average, as attested by the 4.34 normalized run average (NRA)). He finished his career with 13.6 WARP3.

Success #1 is Ubaldo Jimenez., which means there is one legitimate star player that Met fans can point to when they talk about Zach Wheeler’s potential.