From the Desk of the Commish

I have spent the vast majority of this year shaking my head back and forth and throwing my hands in the air in disgust. My wife has seen this expression so many times, she doesn't even have to ask me how my latest series is going; she simply steers clear for awhile until the numbness sets in.

I looked at the team I had assembled last winter and concluded that a division title was a near-certainty. I've been playing this game for a long time, and I like to think I know a division-winning team when I see one. And yet, here I stand 104 games into the season, and not only is my team not running away with the division, but we're four games below .500!

How could I have been so far off in my estimation of my team? Was I simply delusional? Was my judgment so clouded by Cowtippers bias that I couldn't see how much players like Mark Teixeira, Ian Kinsler and Felix Hernandez would suck? Did I not do a thorough-enough job of calculating ballpark factors? Did I not give fielding range enough weight in my forecasts?

As most of you know, I discourage running season simulations using our player disk, as I want to preserve the "mystery" of the regular season as much as possible. However, by this time of the season, I think most of the mystery has disappeared, and we pretty much know how the rest of the season will play out. Last year, I broke this unwritten rule by writing an article for this page that summarized the results of ten simulated seasons. I liked that idea so much, I decided to repeat that exercise this year.

The geniuses responsible for this instrument of torture we call "Diamond Mind Baseball" have stated in the past that if you sim the past MLB season 1,000 times using actual lineups, you will find that the average results correlate nearly perfectly with reality. However, to keep things "fun," they have introduced an element of randomness, so that no simulation is entirely similar to another, and -- as in real life -- anything can happen in any given season. It is this realm of fuzzy randomness that we'll explore today.

For all ten sims, I took our rosters as they appeared on Opening Day (thanks, John Gill!), set all MP pitching rotations, bullpens and lineups, and played out each season with injury ratings. What follows are the results of this study.

Butler	W	L	Pct	Avg RF	Avg RA	pRF	Diff	pRA	Diff	Div	WC	Hi W	Lo W	pW	Diff
Salem	95	65	.594	857	718	794	-63	818	100	8	1	104	87	77	-18
New Milford	89	71	.556	876	777	923	47	811	34	2	5	98	81	93	4
New Hope	68	92	.425	754	881	836	82	922	41	0	0	76	55	70	2
Corona	67	93	.419	781	989	890	109	962	-27	0	0	75	54	67	0

Okay, so I'm not delusional. My team really should be that good. These numbers baffle the mind. The Cowtippers are the only team in the division that has underperformed both offensively AND defensively. And the difference in wins (-18!) is the largest discrepancy in the league. What is least shocking to me is that the Salem pitching staff is on pace to allow ONE HUNDRED RUNS more than they allowed in the average of those ten simulations. Anyone who has had the good fortune of facing Felix Hernandez this season would not be shocked by this. Offensively, we're SIXTY-THREE runs off of the pace set by the ten sims, and that is after adding the likes of Scott Rolen, Ty Wigginton and Kurt Suzuki to the lineup! In ten sims, the Cowtippers never won fewer than 87 games, and yet we're on pace to win ten games fewer than that! In other words, the odds of Salem sucking as hard as they have all season long are less than one-in-ten. If I had the time, I'd love to keep simming seasons to see how many sims it would take for the Cowtippers to win as few as 77 games.

Comparatively, the rest of the division is pretty much performing as expected. It's no surprise that the Blazers are outperforming their average runs scored, but it's only a 5% difference. In ten sims, New Milford won just two division titles, but made it to the playoffs seven times.

New Hope topped out at 76 wins in our ten sims, and they're on pace to win 70 games, so they're right on schedule. And Corona's average sim performance is exactly on target in terms of wins, despite the fact that they're scoring over 100 runs more than their average sim rate. Corona maxed out at 832 runs scored in their best sim, yet they're on pace to score 890 runs this season.

Benes	W	L	Pct	Avg RF	Avg RA	pRF	Diff	pRA	Diff	Div	WC	Hi W	Lo W	pW	Diff
Marlboro	89	71	.556	762	652	673	-89	684	32	9	0	114	79	73	-16
Manchester	74	86	.463	700	778	645	-55	815	37	0	0	79	69	62	-12
Las Vegas	69	91	.431	715	812	671	-44	754	-58	1	0	81	60	71	2
Ravenswood	67	93	.419	645	772	693	48	760	-12	0	0	74	56	70	3

A couple of really disappointing teams at the top of the standings, which explains why the Benes Division race hasn't been much of a "race" all season. The Hammerheads won as many as 114 games in one of the ten sims, but the 35-win difference between highest and lowest number of wins is the largest discrepancy in the league (tied with Sylmar.) In the real world today, Marlboro is on pace to win fewer games than the worst of their ten sim performances. The reason for that, of course, is the white flag Nic Weiss raised early in the season, which continues to wave (I think?) despite his team's first-place standing. That 114-win simulated season was as much of an anomaly as I found in any of the ten sims, as the next-best wins total for Marlboro in any of the other nine sims was just 91. For the most part, the Hammerheads hovered between 83-89 wins, which was good enough to win this weak division nine times out of ten.

For those of us who expected bigger things from the Irish Rebels this season, their 12-win difference between reality and average sim performance is no surprise. Like Salem, Manchester has underperformed both offensively and defensively (no easy trick, trust me.) They're currently on pace to lose seven more games than they lost in any of the ten sims.

Vegas managed to win one division title in ten sims despite a mediocre 81-79 record in that season, and despite being outscored by 46 runs. In ten sims, the Flamingos never outscored their competition, and their runs margin ranged from 46 to 156. They, too, showed a wide range of random outcomes, as they finished two games over .500 in one sim and lost 100 games in another.

Finally, Ravenswood's performance validates Brian Potrafka's early assessment of his team. The Infidels lost 104 games in one sim, and never broke the .500 mark in any of the ten sims.

Yes, you are reading that correctly: the FEWEST number of wins by the Undertakers in any of the ten sims was 104. Yikes. Of course, that 117-win sim would set a new BDBL record. And remember, this was before Paulson added Chipper Jones, Grant Balfour, Justin Duchscherer and others. Had he done nothing to his roster, Paulson could still expect to win 111 games on average with his original team. Absolutely mind-numbing.

But numbing the mind even more is the performance of the Broncs, who have been nipping at the Undertakers' heels all season. San Antonio is currently on pace to win 22 more games than their ten-game sim average, and eight more games than they managed in any one of those sims. They're doing this mostly because they're scoring 121 more runs than they scored in their ten-sim average. Of course, adding Alex Rodriguez to the lineup probably made a difference there.

Sylmar's performance this season seems on par with expectations, with the exception of that 95-win outlier in sim #2. In the other nine sims, Sylmar finished below .500 seven times. That 95-win season was made possible thanks to the performances of bullpen mates Jose Arrendondo (7-0, 1.74 ERA in 114 IP) and Francisco Rodriguez (12-2, 1.90 ERA in 94+ IP), and a ridiculous season by Greg Dobbs (.323/.354/.546, 25 HR), who somehow avoided injury and racked up 520 at-bats. (There is no way to limit playing time in sims, other than by injury.)

Bear Country's performance is exactly in the middle of the upper and lower ranges of their records in the ten sims. There's nothing wrong with consistency -- unless it's consistently sucky.

For all the hand-wringing going on in Allentown all season, you'd expect to see a better performance in the sims. But the table above shows that the Ridgebacks have performed roughly as well as expected this season. Allentown has played .500 ball (as of press time) this season, and averaged 82 wins in our ten sims. Both offensively and defensively, Allentown is performing a little better in "real life" than they did on average in the ten sims. But, although the Ridgebacks captured the EL wild card in 40% of our sims, it looks like our current season will fall within that other 60%.

The Law Dogs' pitching staff is on pace to allow nearly 100 runs more in real life than they did in their ten-sim average. They are eight wins under their ten-sim average, and two wins below their worst performance in ten sims. And yet they sit with a six-game lead in the division at press time.

Both the Mustangs and Sphinx have severely underperformed this season, compared to their ten-sim averages. Both teams are currently on pace to win two games fewer than their worst performance in our ten sims. Great Lakes did manage to capture the EL wild card in one of the ten sims, thanks mostly to an offense that scored 789 runs (62 runs more than any other sim.) Everything seemed to click for the Sphinx in that sim, as Mike Lowell hit .297/.364/.512, J.D. Drew hit .289/.445/.542 (.380/.551/.840 vs. lefties!), Ryan Doumit hit .321/.360/.492, Shane Victorino hit .293/.351/.451 and Cody Ross hit .276/.331/.546. It says a lot, however, that even with everything clicking into place, and even with so many players enjoying outlier seasons, the Sphinx still ranked only fifth in the EL in runs scored.

That 22-win difference between reality and sim average for Southern Cal is by far the largest difference in the Eck League (and tied with San Antonio for the BDBL lead.) The Slyme are currently on pace to win 103 games this season, yet they managed 90 wins only once in ten simulations. The SoCal offense is on pace to score 130 runs more than their ten-sim average, and their pitching staff is on pace to allow 76 runs fewer than that average. How is this possible? I've been trying to figure that out all season. Suffice it to say that Ryan Zimmerman didn't come close to hitting .332/.387/.531 (his current BDBL average) in any of the ten sims. And Torii Hunter's current .333/.402/.556 line is another drastic outlier. The Baseball Gods have been shining on Southern Cal for two years in a row now. Bob must have sacrificed a lot of chicken nuggets to make that happen.

The Apostles are by far the most dominant Eck League team in our ten sims. The 900 runs scored average is not surprising at all, but that 648 average runs allowed is difficult to fathom. I must be seriously underestimating the awesomeness of Tim Wakefield, Randy Wolf, Jose Contreras and Kyle Lohse. Their performance in the "real" BDBL season seems much more realistic to me, yet they are on pace to allow 181 runs more than their ten-sim average! That's the largest difference in the league, and 81 runs more than Salem's!

Nashville is underperforming across the board, though some of that certainly has been caused by trades. And South Carolina is performing about as well as could be expected (which is to say, not well at all.)

Both the Black Sox (who are a game ahead in first place at press time) and Rocks (who are in last place despite a record better than .500) are outperforming their ten-sim averages by more than 15 games. The Chicago offense is on pace to score 117 runs more than their average, and their projected total of 91 wins is ten more than their best performance in any of the ten sims.

While Atlanta is within their ten-sim wins range (89-103), their 90-win pace has to be considered something of a letdown, given that they won this division in all ten of the simmed seasons, and were the only team to average more than 75 wins. The Fire Ants' offense has underperformed relative to their ten-sim average, while the pitching staff has outperformed that average (and is second only to the ridiculous Undertakers.)

Obviously, there is a significant amount of randomness in this game we play, not only in a short series (which we all expect), but over a large, 160-game sample size as well. It is this latter randomness that I think we tend to underestimate. The standard deviation for wins (i.e. the average difference in wins from one season to another) in our ten sims is a little more than six. Think about how many division titles (and wild cards) have been won by fewer than six games, and that will give you some idea as to the importance of randomness in this game that we play.

While this randomness makes this game so frustrating, you feel like throwing your fist through a cement wall, it is a necessary evil. If it weren't for random fluctuation, there would be no reason whatsoever to play out the season. Just do a few calculations at the start of the season and award the trophy to the best team.

It is also worth mentioning that a lot of the large discrepancies above are self-perpetuating. Depending on a team's performance over a much smaller sample (the first 28-56 games of the season), a team may decide to "go for it" or "pack it in," thus completely changing the pre-season expectations of the team and compounding the good or bad fortune of the first chapter or two.

While it may not help me sleep better at night knowing my team should be much better than they are, it is a tiny bit comforting to know that I am not delusional, and that I can stop questioning everything I know about the game of baseball.

Here's How It Could Have Played Out