The stats projections for the 2023 season are now up. Actually they first went up around Christmas, but I’ve been doing an unusual amount of tinkering this time around. Updates should happen at least a couple of times a week between now and the start of the season – which kicks off on March 30 this year.

1. Rules changes for 2023.

Trying to figure out how the reduced shifting and the new pickoff attempt rules will affect the statistics was definitely an exercise. A lot of it, admittedly, is guesswork, but we do have the results from the minor leagues to guide us. The gist of it is I raised stolen base attempts for all players by 25% (which obviously makes a bigger difference for guys who steal 20 bases than guys who steal 2), while decreasing caught stealing rates. Left-handed hitters received improved single rates, enough to raise their batting averages by about 10 points. Switch-hitters got a bout a 7-point raise. Righties get nothing. And similar adjustments were made for pitchers, with right-handed pitchers (who face more lefties) getting a slightly larger downgrade on their BABIP.

2. Pitcher adjustments

Every year, I make adjustments made to the way I translate between the player statistics and what I expect in the next season. The changes in the pitcher numbers are somewhat larger this year. I made a mistake last year, which did not become fully obvious to me until later, in that I way over-did the regression to mean aspect. All pitchers, good, bad, or otherwise, got squeezed into a tight bin, with not nearly enough variation. I tried to avoid that trap this year.

3. Stronger adjustments for foreign leagues

There are some oddities in projecting statistics from Japan, and I took a deep dive into them this year after a disappointing performance from some more disappointments. I’ve always calculated league strength by using the results of all players who played in those leagues. I had recognized a problem with that before – that the translations between AAA and Japan gave very different results than those between the majors and Japan – but extended it this year to noting where players are from. The result of that is Japanese hitters losing about 20 points of EqA in their translation. I will try to drw that out in another post.

4. Fielding changes

A friend of mine pointed out that the fielding numbers in the translation seemed off. This was something that I think I’ve known for a while. I had built in a sort of universal fielding adjustment to handle “defense” without having to track the position closely, and it just didn’t work. So I totally overhauled the way that worked, which should lead to more consistent values.

5. New schedules

The first set of numbers I put out still had the 2022 schedule built into it. Most years that is no big deal, but this year they are changing to a more balanced schedule, with less concentration on your own division. That was good for teams in the strong Eastern divisions (the Nationals gained 0.77 wins when I changed the schedules, most of any team; the Marlins, Red Sox, and Orioles all gained over 0.5 wins). And it was bad for the AL Central (the Guardians and White Sox were the two biggest losers, dropping by 0.85 and 0.69 wins). Those changes in team totals will be accompanied by changes in individual numbers, but it isn’t large enough to be a real issue.

6. org Files with Roster Status

One of my personal favorite tools while gearing up for the season are the org dt pages. Each team is broken down by position (with pitchers grouped into Starters, Relievers, and Swing), and each player is ranked in order of their projected Major League playing time. I find it a good way to see a team’s entire depth chart – including their coming league players. And I now have it cross-referencing with MLB roster status, so you can see who is on the 40-man roster (“A”, until we get into the season), who is an NRI (“n), and who is injured (“D”, when we get that far). Here are Brewer shortstops:

    SS
  Last         First       Team Lg  Age  PA AB   R    H  2B  3B  HR RBI  BB  SO  SB  CS   BA    OBP   SLG  EqBA EqOBP EqSLG  EqA   VORP  WARP Defense FRAA  MJ BRK IMP CLP ATT DRP UPS 
A Adames       Willy        MIL NL  27  597 543  81 147  28   2  28  89  53 170   7   4 0.271 0.337 0.484 0.272 0.340 0.488 0.284  43.1   5.2 140-SS   4    94  10  42  35  18   8 133 
n Alvarez      Eddy         MIL NL  33  433 386  50  99  16   2  10  39  35 117   7   2 0.256 0.337 0.386 0.266 0.345 0.394 0.264  20.6   2.2 100-SS  -1    61  16  36  56  26   8  20 
A Turang       Brice        MIL NL  23  625 576  69 152  25   3  10  60  49 143  22   3 0.264 0.322 0.370 0.264 0.324 0.363 0.256  25.1   3.4 147-SS   5    40  27  57  30   8   1  54 
n Monasterio   Andruw       MIL NL  26  437 398  53  97  18   2   9  43  35 116   9   2 0.244 0.311 0.367 0.258 0.326 0.375 0.249  14.1   1.3 102-SS  -3    34  23  56  48  25   6  33 
  Devanney     Cam          MIL NL  26  587 539  59 117  26   2  17  59  41 165   5   2 0.217 0.281 0.367 0.231 0.294 0.376 0.233  10.1   0.9 137-SS  -2    24  31  52  41  24   5  32 
  Brown        Eric         MIL NL  22  305 278  33  58  14   2   6  23  22  81  19   2 0.209 0.279 0.338 0.226 0.295 0.343 0.237   6.4   0.8  71-SS   1    15  23  40  40   1   0  50 
  Zamora       Freddy       MIL NL  24  231 218  20  46   6   1   3  13  10  63   6   1 0.211 0.255 0.289 0.211 0.254 0.284 0.202  -2.7  -0.4  54-SS  -1     5  64  74  27  20   1   5 
  Garcia       Eduardo      MIL NL  20  541 516  52 110  22   2  13  53  20 217  12   1 0.213 0.250 0.339 0.227 0.263 0.350 0.214  -0.5  -0.3 127-SS  -2     2  43  65  30   8   4  15 
  Murray       Ethan        MIL NL  23  427 393  38  79  16   2   8  35  31 137   9   2 0.201 0.265 0.313 0.215 0.279 0.315 0.214  -0.7   0.5 100-SS   5     2  28  51  37  13   2  11 
  Barrios      Gregory      MIL NL  19  363 341  36  78  13   2   1  28  19  82  13   1 0.229 0.275 0.287 0.233 0.279 0.289 0.214  -0.4   0.9  85-SS   8     0  61  76  22   6   1  13 
  Guilarte     Daniel       MIL NL  19  266 246  20  54   9   1   1  21  20  79   9   2 0.220 0.278 0.276 0.228 0.288 0.281 0.211  -1.0  -0.3  63-SS  -2     0  49  62  27   3   0  13 
 

I don’t know more than anyone else when games might start being played…but I do have my first look at team projections here.

Just like last year, it looks like the Dodgers and Padres are the best two teams. The Padres collapsed, epically; the Dodgers made it to the LCS before losing to the (eventual WS champ) Braves.

 

My spare time isn’t what it used to be…NOAA keeps me busy. But I finally do have the new MLB pbp files parsed enough to run all the split data I need. So all of the player, league, and team should be back to having those breakdowns. Not a fan of MLB’s chosen format, although to be fair I am not using their chosen tools for reading it out. It is really, really non-linear; I think my code has to make 5 separate passes through a game to get everything it needs and arranged in a game order.

Cool graph today is the playoff picture of the AL East:



The Rays (green) remain steady at the top. The Yankee (black) line surges through their August winning streak, but has been collapsing since. The Blue Jays track (blue) over the last month is a mirror image of the Yankees, with the Red Sox (red) line running right between both of them. There is a little better than a 4-in-5 chance that both AL wild cards come from the East…two out of three teams that right now are basically even.

Also in playoff odds…the Giants this morning became the first team to make the playoffs in all 1 million runs, giving them a 100% score. The Dodgers fell just short; their 99.99995 means they finished playoff tie in 1 out of a million runs, and made the playoffs in the other 999,999 runs.

 

The “DTs by league” pages are not updating right now, because I can’t get the play-by-play data I use to build them from MLB. Trying to figure out what’s gone wrong in my connections. The EQA report is mostly fine – except for splits data which, once again, requires the PBP files.

 
These come from my million-run Postseason Odds page, which will update daily through the season.

Average wins by position in AL East:  100.4 91.4 84.8 77.4 58.1 
AL East    Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
Yankees    .593   97.5   64.5   63.58508   23.74102   87.32610 
Blue Jays  .529   86.1   75.9   13.51403   31.84269   45.35672 
Rays       .525   85.9   76.1   13.07292   31.31225   44.38516 
Red Sox    .518   84.2   77.8    9.81512   27.01932   36.83444 
Orioles    .370   58.4  103.6     .01285     .10404     .11689 

The Yankees are an overwhelming favorite to win the division. The 
Orioles have the lowest playoff odds of any team in the majors. At 
least one of the remaining teams should take a wild card, and 
there's a pretty good chance that both wild cards come from the 
East.

Average wins by position in AL Central:  94.7 86.5 79.7 72.6 63.6 
AL Central Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
White Sox  .528   88.7   73.3   43.41552   16.64380   60.05932 
Twins      .527   88.1   73.9   40.12318   17.02435   57.14754 
Clevelands .478   79.6   82.4   11.56375    9.88443   21.44818 
Royals     .444   73.7   88.3    3.96877    3.91370    7.88246 
Tigers     .406   67.1   94.9     .92878     .95012    1.87890 

The Twins and White Sox are rated pretty close to even, with me 
giving the barest edge to the White Sox. The ex-Sockalexises have a 
chance, if their pitching can carry them through. The Royals have 
their best chance in several years, and some prospects whispering 
hopes of even better times ahead. I expect the Tigers to challenge 
the Orioles for the barrel bottom.

Average wins by position in AL west:  96.7 85.3 78.9 73.1 65.7 
AL West    Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
Astros     .576   95.2   66.8   77.30096    7.74413   85.04509 
Athletics  .489   80.0   82.0   10.00943   12.37409   22.38351 
Angels     .477   77.4   84.6    6.30802    8.42805   14.73607 
Mariners   .461   75.2   86.8    4.22883    5.91044   10.13927 
Rangers    .441   71.9   90.1    2.15278    3.10758    5.26035 

The Astros get rated as the single most likely division champion, 
thanks to their own strengths and contenders' weaknesses. 

Average wins by AL First Wild Card:  93.1 
Average wins by AL Second Wild Card:  88.8 

Average wins by position in NL East:  95.8 88.5 82.5 76.0 65.3 
NL East    Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
Mets       .557   88.7   73.3   36.46963   17.94029   54.40992 
Braves     .554   88.3   73.7   34.61201   17.93041   52.54242 
Nationals  .528   83.9   78.1   18.43439   14.48495   32.91933 
Phillies   .507   80.1   81.9    9.84990    9.46676   19.31666 
Marlins    .435   67.1   94.9     .63407     .78651    1.42058 

A highly contentious division, with almost everyone having a chance. 

Average wins by position in NL Central:  93.3 86.1 80.3 73.7 61.2 
NL Central Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
Brewers    .529   86.1   75.9   36.00577   10.24474   46.25051 
Cardinals  .520   84.6   77.4   29.57912   10.05231   39.63143 
Reds       .507   82.3   79.7   21.18838    8.59095   29.77934 
Cubs       .491   79.2   82.8   12.87463    6.10106   18.97569 
Pirates    .394   62.4   99.6     .35211     .16244     .51455 

An even more contentious division, although the teams are all 1-5 
games behind the East group. It is the weakest division.

Average wins by position in NL West:  107.2 96.4 81.5 72.4 60.9 
NL West    Pct3  Avg W  Avg L  Champions  Wild Card   Playoffs 
Dodgers    .646  104.0   58.0   63.67985   32.50205   96.18190 
Padres     .620   99.1   62.9   35.21897   54.38710   89.60607 
Dbacks     .498   77.9   84.1     .75033   11.05648   11.80681 
Giants     .480   74.8   87.2     .34225    6.02418    6.36643 
Rockies    .412   62.8   99.2     .00860     .26977     .27837 

What seems very likely is that the Dodgers and Padres - with the two 
best projected records in the majors - will both advance to the 
playoffs, leaving the rest of the NL to fight over the remaining WC 
spot. 

Average wins by NL First Wild Card:  97.4 
Average wins by NL Second Wild Card:  90.0

 

Although things don’t look much different, if we dive down under the hood you will be able to see that I have been a lot of work in the background. And I guess one of the best ways to demonstrate is to look at this page here:

http://www.claydavenport.com/ht/TORKELSON19990826A.shtml

I am now able to provide a stat-based projection on 2020’s #1 overall pick, Spencer Torkelson, based off of his college statistics. Without them I had literally nothing on Torkelson to base any stats. With them – that is a projection that makes him a very legitimate callup.

And he is not alone. Adley Rutschmann jumps from a .240 eqa projection to .284 – on the basis of his college pedigree. Andrew Vaughn likewise goes from the mid-230s to the mid-270s – because now I know he had a pair of .300 translated seasons in college.

I’m still early in my use of college stats – haven’t gotten fielding info in yet, which is going to make bring Rutschmann’s drop once it knows he’s a catcher – but it seems to work as well as minor league numbers, with a couple of caveats.

I haven’t backed these up with a formal study yet, but it certainly looks like there is a very tendency for a player’s final year in college to be a lot better than his prior years. Rutschmann and Torkelson both fit that pattern. I think that pattern will be especially pronounced for 2020 – play was interrupted typically before in-conference games started, meaning that teams (from the major conferences, at least) played unusually weak schedules. The way that I calculate the difficulties, averaging across different years, is going to mask that weakness.

The second caveat is that players appear to underperform their college numbers in their first taste of pro ball, but tend to recover towards their college averages in later seasons. There is this double whammy of a player’s final college year coming in too hot, and his first pro year too cold, that creates a LOT of distortion.

About the difficulty ratings. So the ratings I am using here see the top three college conferences as the Southeastern, the Pac-12, and the ACC. This is in line with the number of players who have gone on from these conferences to play in Organized Baseball. I counted 761 SEC players in OB ranks since 2011, 611 ACC players, and 565 from the Pac-12. Their ratings, of .498, .458, and .478 are comparable to the short-season Northwest (.469) and New York-Penn (.514) leagues (all of these numbers area average from 2015-2019). Your Midwest and Sally leagues in the .55-.57 range, the high A leagues are .63 give or take, with the majors sitting at just above 1.00 right now.

Here’s how the top 16 conferences, ranked by number of OB players, rank:

Conference       Players Diff
             -->  NY-Penn   .514
 Southeastern      761 0.498
 Atlantic Coast    611 0.458
 Pacific 12        565 0.478
             -->  Northwest .469
             -->  Pioneer   .464
             -->  Appalachian .462
 Big 12            464 0.432
 Conference USA    364 0.443
 Big West          347 0.433
 Big 10            340 0.415
             --> Gulf Coast  .423
             --> Arizona     .418
 Sun Belt          274 0.407
 West Coast        263 0.420
 Southland         258 0.388
 Big East          266 0.432
 Western Athletic  242 0.372
 Missouri Valley   219 0.432
 Mountain West     217 0.435
 American Athletic 214 0.413
 Atlantic Sun      205 0.411

There is a .81 correlation between the number of players taken from each league and its quality, so the draft appears to be resosnably efficient The ratings do tend to keep dropping from there. By the time you get to, say, the Wisconsin Intercollegiate League, made up of the smaller schools of the University of Wisconsin system (like Oshkosh, Stevens Point, or Whitewater), the rating is coming in at just .130.

Hat tip to baseballreference.com, who really pulled these collegiate numbers together. My task was just to pull them together in a sensible fashion

 




Major League Baseball has never had a player whose last name starts with X. Joe Xavier, who made it to AAA in 1989-90, is as close as any X player has gotten.

Until now.

Major League Baseball’s decision to reclassify some of the Negro Leagues  as “major league” changes that.  The Negro American League – once of the reclassified leagues – had a man by the name of Leovigildo Xiques, who played for the (Cincinnati and) Indianapolis Clowns.


 

The title says it all. They are under “DTs by League“.

 

First things first – I now have the way too soon, already out of date, Team Projections for 2021 up and available.

Of course, posting them on the day of the non-tender deadline means most teams have at least one player in my projection that is already compromised. And of course, teams have yet to fill the gaping holes (like the Nationals 1B slot) that you know will be taken by some current free agent. Still, its nice to have a snapshot of where the team stood when they started making decisions about 2021.

generated on 12- 2-2020, projections for 2021 season

AL East Won Lost Runs RunsA
NYY 92 70 845 737
TBY 86 76 769 726
BOS 83 79 837 816
TOR 78 84 779 803
BAL 70 92 707 808
AL Cent Won Lost Runs RunsA
MIN 87 75 809 747
CLE 84 78 764 749
CWS 82 80 788 772
KCR 72 90 714 799
DET 69 93 673 785
AL West Won Lost Runs RunsA
HOU 89 73 818 742
LAA 83 79 812 792
OAK 81 81 743 743
TEX 77 85 755 783
SEA 71 91 703 791
NL East Won Lost Runs RunsA
ATL 91 71 815 719
NYM 86 76 752 707
WAS 83 79 735 723
PHI 80 82 738 750
MIA 74 88 673 739
NL Cent Won Lost Runs RunsA
MIL 86 76 750 706
CHC 83 79 756 738
CIN 82 80 755 748
STL 77 85 707 744
PIT 73 89 700 777
NL West Won Lost Runs RunsA
LAD 96 66 843 704
SDP 89 73 779 707
ARI 81 81 768 768
COL 77 85 783 819
SFG 67 95 665 792

One thing I am pretty certain about – not just for me, but everybody else doing projections, from BP to FanGraphs to ESPN, Rotowire, Rotoworld, and every other fantasy site – projections are going to be worse in 2021 than in prior years. Having a 2020 season that was only 60 games – and even then, only for the majors – every player is coming off of what we would, previously, have labelled a missed or truncated season. Our sample sizes are smaller than ever.

To balance that, I tried something I never did before. The “2020” season that is going through the program, combining with 2019 and 2018 stats to make a 2021 projections, is a blend of whatever the player really did in 2020 with the projection I made going in to 2020. For every minor leaguer, who had no 2020 season at all (or at least nothing with public statistics!), there is nothing to blend – the 2020 projection is their entire line.

I can’t say I’m enthused with the results, but testing says it was a lot better than going with zero seasons or taking 2020 results at full face value.

 

Welcome to anyone who might be checking in on me for the first time. Or for the first time in a long time. Either way is good for me.

As you can plainly see, I don’t write that often. At heart, I am a number-cruncher, not a wordsmith, and this site is about the numbers that are behind the tabs, not the words I use to highlight or stretch those numbers. And that in turn is because I am not thinking of this as a public site; it is really a vessel for me to have a lot of this work available, to me,  when I am not sitting at my own desk.

So a quick guide to whats on here…

The EQA Report tab is for the major leagues, only. I have tables for all major league seasons, back to 1871, which update every day in season. The centerpiece of these tables is a statistic I called Equivalent Average, which measures a players total offensive value and scales it to the historical range of batting average. Not because batting average is superior in any way – it certainly is not – but because baseball fans, at least of my generation, have an immediate grasp of the scale.

Next up is the Playoff Chances tab. This is the output of a million-run Monte Carlo simulation of the rest of the season – at this point, the entire season.  My version has a wider spread than similar charts run by other sites; I believe they are too settled on their valuation of (say, the Yankees) as a “.600 team”.  I use some tricks to build a distribution of values for the Yankees. In most runs, they are a .600. But in some runs they’re a .650 team; in others, and more, because it is easier for things to go wrong than right, they only get treated as  a .550 team. That’s how the chances for even the Orioles, while miniscule, are not zero.

Then is DTs By League. These are statistics for minor league players, organized by league. There are three flavors to these pages. The first flavor is the real stats  – completely unadjusted. The second flavor is the DT – the Davenport Translation, originally developed in the late 80s – which estimates what that the real statistical line is worth, right now, in major league terms. The third flavor – peak DT – builds on the second flavor but adds “normal” improvement with age. For the US minor leagues, I have these stats going back to 1979. Triple-A leagues go back to 1946. Japanese leagues go into the 1930s. Other foreign, winter, and independent leagues are randomly available.

DTS by Organization are probably the pages I myself look at most often during the season. It is only going to have players for the current season, arranged by major league team, with all of their minor league affiliates listed with them.

The Projected Standings is where projected statistics, for team and players, resides. The projections have two flavors to them. One is the straight out of the computer numbers; the other is the result of loading those numbers into a very large spreadsheet, manually assigning playing time to everyone, and then rebalancing the numbers so that the hitting and pitching totals are consistent. Those manual adjustments are a real chore, and difficult for me to maintain through the season.

The last tab is the Spring Stats, which is going to start populating Any Day Now. While they are generally of limited predictive value, they certainly do affect who actually makes it out of spring training. There is also a second tab in there, for Current Team DTs, which has three entries per team – all of the team’s hitters (“.2020dt”), all of their pitchers (“.2020pdt”), and a depth chart (“.2020orgdt”) culled from the computer-only projections.

 

 

 
Set your Twitter account name in your settings to use the TwitterBar Section.