Diamond Mind Projection Blowout

[haltz]

Today, VEB tipped off readers to the DM sims run by a site called Replacement Level Yankees Weblog, featuring four different projection systems. The entry is way too lengthy to post here, but I'd recommend clicking on the link, there is a lot of interesting stuff there. Here's lboros' abbreviated version:

on tuesday night, SG posted the results of the 3d annual Diamond Mind projection blowout at Replacement Level Yankees Weblog. he ran the 2007 season 4,000 times, using four different sets of projection data (chone, PECOTA, ZIPS, and Diamond Mind) for 1,000 seasons apiece. head over to SG's site for all the details, which are plenty juicy; i'll just post the aggregated results (ie, the average of all 4,000 seasons) for the nl central:

Code: Select all

			          W 	 L 	GB 	 RS 	 RA
st louis 			84 	78 	-- 	758 	721
chicago 			 84 	78 	-- 	808 	775
milwaukee 		  81 	81 	 3 	764 	768
houston 			 80 	82 	 4 	775 	788
cincinnati 	    73 	89 	11 	749 	819
pittsburgh 	    73 	89 	11 	723 	798

actually, stl edged out the cubs by .2 of a game, 84.4 wins to 84.2. in other words, over those 4,000 seasons the cardinals won about 337,600 games, while the cubs won about 336,800. that's pretty tight, boys and girls. a few other items of note:

* the cards' worst showing came in the PECOTA projection --- they averaged an 80-82 record, fourth best in the division and 5 games worse than the cubs.
* their best showing was in the ZIPS projection --- average record of 90-72, the best not only in the division but in the whole national league, and second only to the yankees in mlb.
* they finished no worse than 2d in the national league in runs allowed in any of the four projections. in the aggregate projection (all 4,000 seasons) they averaged the lowest runs-allowed total in baseball, 721.
* among the four projections, the cardinals finished no better than 9th (tied) in the national league in runs scored --- and were as low as 15th (PECOTA); in the aggregate, the offense ranks 13th out of 16 national league clubs. for all the worry about the pitching staff, the simulation exercises have consistently indicated that the offense is where our concern ought to lie. the offense could really suck.
* the cardinals won their division 1,382 times, or 35 percent of the time --- 3d-best showing in the national league (the phillies and padres won their divisions 39 pct of the time). the cards made the playoffs as either a division winner or wild card 41.6 percent of the time, 4th best showing in the league (behind the mets, padres, and phillies).

[Popeye_Card]

Wow. I'm quite surprised about those pitching numbers. Especially since Wells' and Reyes' projections don't look very good. I wonder if part of it is that they took Looper's projections that were made as a reliever, and didn't make any corrections for being a starter? I'm sure a lot of teams would benefit from taking a reasonably good reliever ERA (Looper ranged from 3.53 to 4.10) and getting 200 innings out of it.

--P--

[GatewaySnayke]

LINK

FWIW, the Cardinals and Cubs posted I believe an aggregate of 84-78.

St. Louis had a very weak offense, and I don't think 721 RA is anything to be proud of.

Chicago had 808 RS and 775 RA.

[Transmogrified Tiger]

hahaha, four THOUSAND years of simulations and the Royals can't make the playoffs once

[Popeye_Card]

and I don't think 721 RA is anything to be proud of.

According the the notes in this thread, that's the lowest average in MLB.

--P--

[Popeye_Card]

hahaha, four THOUSAND years of simulations and the Royals can't make the playoffs once

That's not as surprising to me as the White Sox only making the playoffs 119 times (3.0%). That's almost as low as the Pirates (117 times). I realize they play in a tougher division, but that's rough.

--P--

[GatewaySnayke]

and I don't think 721 RA is anything to be proud of.

According the the notes in this thread, that's the lowest average in MLB.

--P--

Which doesn't make sense to me.

If 721 is the lowest average, wouldn't that indicate we are in a BIG offensive boom period? Why aren't the Cardinals reaping the rewards of this boom then?

[Popeye_Card]

and I don't think 721 RA is anything to be proud of.

According the the notes in this thread, that's the lowest average in MLB.

--P--

Which doesn't make sense to me.

If 721 is the lowest average, wouldn't that indicate we are in a BIG offensive boom period? Why aren't the Cardinals reaping the rewards of this boom then?

No, it's the lowest average. 2-3 teams in each league will probably best that number (including the Cards, whose lowest was 670--lowest in MLB, and would have led MLB last year too). But the average for each team is going to be higher.

--P--

[jim]

Here are his projections with the DMB disk compared to mine. Ours came out pretty much the same. Mine is on the left.

Code: Select all

AL East
NYA   96.20/97.4  
TOR   90.48/89.1  
BOS   85.30/86.5  
BAL   75.76/74.9  
TB    69.20/69.3  

AL Central
DET   90.26/88.8  
CLE   89.90/91.5  
MIN   88.4/88.1   
CHA   76.10/77.4  
KC    64.32/65  

AL West
LAA   85.00/86.4  
OAK   83.96/83.8  
SEA   78.62/77.8  
TEX   78.20/77  

NL East
PHI   85.84/84.7  
ATL   85.08/84.6  
NYN   82.08/82.5  
WAS   74.74/75.4  
FLA   73.06/73.3  

NL Central
STL   84.16/83.5  
CHN   83.38/82.5  
MIL   79.48/79.6  
HOU   79.06/81.4  
CIN   76.24/75.8  
PIT   72.12/72.1  

NL West
SD    87.80/88  
LAN   83.16/82.4  
ARI   78.06/78  
SF    76.90/76.7  
COL   76.86/76.6  

[haltz]

Wow. I'm quite surprised about those pitching numbers. Especially since Wells' and Reyes' projections don't look very good. I wonder if part of it is that they took Looper's projections that were made as a reliever, and didn't make any corrections for being a starter? I'm sure a lot of teams would benefit from taking a reasonably good reliever ERA (Looper ranged from 3.53 to 4.10) and getting 200 innings out of it.

I have no idea about the Diamond Mind projections. I also don't know enough about the DM software itself, having never used it, to know what would happen if you plugged his reliever numbers into a starting role (or what SG at RLYW did concerning this). CHONE projected him as a reliever, PECOTA as a starter, and Dan Szymborski went to the trouble of projecting him in both roles for ZiPS, not sure which one or if both went on the build however.

By ERA, the projections for our starting five look like this according to PECOTA, CHONE, and ZiPS:

Code: Select all

                 PECOTA        ZiPS       CHONE
Carpenter         3.26         3.06       3.13
Reyes             3.81         3.86       3.37
Wainwright        3.88         3.89       3.23* 
Wells             4.47         4.98       4.26 
Looper            4.82         4.72       3.73* 

*reliever