Another set of simulations using projections
DMZ · March 24, 2008 at 2:22 pm · Filed Under General baseball
From the Replacement Level Yankees Weblog: using DMB and the CHONE projections, 1,000 seasons of Mariner baseball came out with the M’s finishing 83-79 and making the playoffs 20% of the time.
Many caveats and disclaimers apply, as with all these kind of things — check out the article for more.
Comments
28 Responses to “Another set of simulations using projections”
Leave a Reply
You must be logged in to post a comment.
The Devil Rays have a 30% shot at the playoffs. I knew they were on the upswing but, wow. That’s interesting.
I find it funny how many local beat writers talk about how we’ve caught the Angels . . . looks to me like the angels have the highest chance of making the playoffs out of any team in MLB by what that sim says.
My take away from this is that our chances of winning the division beat our chances at the wild card by a factor of 3-1. Confirming what we already knew. We did win the division 153 times, which beats 73 times from 07, 54 times from 06 and 110 times from 05.
So at least as far as the projections are concrned this is the best team we fielded in years. I can’t see any reason to disagree with that.
I’m not 100% positive, but I don’t believe those projections take into account the injuries to the Angels’ players. So as much as it may seem like a disappointment to M’s fans, I think we’re slightly better off than those specific projections expect. Though clearly not by much.
Also, PECOTA and ZiPS hate us because they are not fond of Ichiro’s skillset, among other things, correct?
It’s a good thing they actually play the games.
I find it funny how many local beat writers talk about how we’ve caught the Angels . . . looks to me like the angels have the highest chance of making the playoffs out of any team in MLB by what that sim says.
The Angels’ odds of making the playoffs, relative to most other MLB teams, are inherently better. They only have to compete with three other teams, as opposed to four or even five, in order to win the division and make the playoffs. When it’s generally accepted that two of those three can be written off, that only strengthens the argument. So comparing their odds to the field at large isn’t that germane to how close a rival within the division might be to catching them.
The same principle applies to the Mariners winning the division over the wild card. The number of teams you have to beat out is a huge factor. It’s why you’d just as soon be seven games back, with but one team in front of you, than five games back with three teams between you and a playoff spot.
I’m not 100% positive, but I don’t believe those projections take into account the injuries to the Angels’ players.
I did factor in the Escobar and Lackey injuries. I assumed Escobar would miss 20% of the season and Lackey 15%.
Don’t take these too seriously. I projected the 2005 White Sox to win 79 games after all…
#6
All great points. But just saying that all this talk of us being contenders, and Bedard bringing us up to LAA’s level, etc is still quite overblown. I want us to make the playoffs, but at the same time, if we tank and that is what it takes for Lincoln et al to be shown the door . . .
Don’t take these too seriously. I projected the 2005 White Sox to win 79 games after all…
That a few outliers exist isn’t surprising–in fact it’s to be expected. You’re pretty clear about acknowledging that there is variance around the mean. There’s really nothing wrong with being far off on a few teams as long as there aren’t any underlying reasons that those teams were incorrectly projected.
It looks like if the Mariners play 1 SD above their expected win total they are at 84 wins and if the Angels play 1 SD below their expected win total, they are at 82 wins. Unless someone can point out a material reason that there is a bias in the system–and give evidence that the bias actually exists–then it seems fairly clear that we should both expect the Angels to win the division but not completely count out the Mariners.
See, that’s a problem and what so many of us having screaming about all off-season. The premise of the Bedard trade was the the M’s were within a few games of catching and passing the Angels. Clearly (as much as we can take them seriously, as SG said) the M’s aren’t. These systems may have hits and misses (Dan’s ZiPS has the A’s winning a lot more games than any other system) but generally fall inline with the real results.
If we’re only a little bit better than these projections and the Angels only a little worse (even accounting for the injuries already), the M’s are in a heap of trouble.
Also, what I found really interesting was the horrific offensive output of the Mariners. Using all 6 systems (6000 seasons) the M’s average 717 runs a year- 2nd to last in MLB, in front of SF’s 684 only. Ick.
Seems like you’re on the ball SG. Ignore my thoughts.
Huh. Most of those projections are forecasting the Mariners to have one of the worst offenses in the league- around 680-720 runs scored. Basically, they are expecting a complete collapse from last year’s 794 runs (the M’s were actually an above-average offense last year when adjusted for Safeco).
To put that in context, the 2004/2005 Mariners scored about 700 runs both years (698/699), and the Mariners got to that point in both of those years by having the trainwreck of Wilson/Olerud/Spiezio/Aurilia/Edgar with a combined OPS WELL under .700 in 2004, and by having a historically bad peformance at C in 2005, combined with very bad performance from 2B (Boone/Lopez) and CF (Reed).
It seems to me most of the projection systems are projecting collapse in the Johjima/Sexson/Raul/Wilkerson axis of our offense… and since Sexson and Wilkerson are already pretty bad and Raul spent half a year looking awful, that might not be far off the mark.
The projections tend to be very down on Lopez, too, IIRC.
Most of the systems are available for free on fangraphs. PECOTA is on BPro and ZiPS is available for free (thanks Dan!) on Baseball Think Factory.
They’re all pretty interesting and could give each person some insight as to why they think the M’s will be…less than what they think they’ll be. Of course, if you havn’t gotten it yet reading Dave’s and Derek’s post than Nate Silver probably won’t convince you either.
What does that say about Billy Beane if the A’s who are in a serious rebuilding year compete for second with the Mariners who are in a win at all costs year?
These projections are fun to look at, but are years and years away from being close to reliable enough to warrant a statement like #2.
17 again, my point wasn’t to say that the angels are a lead pipe cinch, just saying that its far from being up in the air (50/50) like so many people (mainly in the M’s front office) seem to believe.
These projections are fun to look at, but are years and years away from being close to reliable enough to warrant a statement like #2.
If you look at statistical projections for 2007 and compare them to the results of the 2007 season, the Diamond Mind and PECOTA projections did better at projecting the season than any human analyst that I could find (certainly better than the poll of espn baseball analysts), and Nate Silver made his PECOTA projections worse by trying to adjust for things he thought that PECOTA didn’t capture. I didn’t check SG’s ’07 projection blowout, but I would suspect they were similarly good.
The projections are nearly to the point where they can’t get any better. There’s a certain amount of variance innate to a series of 162 win-loss events, and nothing can predict the future with 100% certainty. But if you’re going to just off-handedly dismiss simulation results, you basically have to dismiss the idea that anyone has any idea which teams are the best. And if things were really that random, the Yankees and Red Sox wouldn’t have half of the last ten World Series titles.
17: Really? Post 2 looks accurate to me. After all — though this point is often lost early in debates about projections — post 2 isn’t making promises, simply noting something evident in the likelihoods.
18–
Must have missed it…who in the front office said we were 50/50 to win the division?
It’s almost as if a team needs to have players who haven’t been measured in the past in order to surpass their projections. For instance, the Colorado Rockies were picked for 79 wins yet made the world series, in part because of the outstanding play of their young players.
Jettison Vidro, Bloomquist, Wilkerson and Sexson and play Balentien, Morse, Clement, LaHair and hope for the best.
And also, there’s the intangibles, the good guy in the clubhouse, the Carlos Guillen factor (the only time the 116 win Mariners had a losing streak was when Guillen was hurt).
50/50 chance I’ll win the lotto too. I’ll either win it, or I won’t.
the only time the 116 win Mariners had a losing streak was when Guillen was hurt
Ooooooh, correlation! I know what comes next!
22 – you mean the Miguel Cairo factor! True veteran grit.
And McLaren was quoted on sending Balentin back down to Triple-A, saying, “He left a real nice impression. He stayed inside the ball, used the whole field and showed some power. I want him working on his defense, try to steal a few bases, and improve on his all-around game.”
Sadly, McLaren has no desire for anyone else in the organization to work on their defense, or else we wouldn’t be sending dead-men-walking Sexson and Ibanez out to field like little league champions.
21- First off, I get the sense that you’re trying to bait me into an argument splitting hairs. I know I am not the only person who has noticed the following pattern:
1. The M’s management sees we won nearly 90 games last year.
2. While our pythagorean points to luck, our front office believes that truly is the baseline for this years team, had we not made any changes.
3. We added Bedard while “only giving up some prospects” meaning that we basically have last years team intact.
4. Take a ~89 win team, add another ace to the mix, and voila, we are a contender for the division.
Point is, while they haven’t explicitly stated it as being 50/50, they are still delusional as to the true talent of this team and its playoff chances.
[violation of button terms]
Seriously? You guys are seriously arguing that because sims can predict better than humans (how about the VOLUMES of info you’d need to prove that), they’re valid? You really are drinking the cool aid… and I’m a huge proponent of sims. Are we factoring in to the sim how the Yankees will make deadline moves to get better halfway through the season, or that Oakland may easily trade the few guys he has left to fleece someone at the trade deadline… etc. etc. My point is that the chaos of the baseball season is so vast that statements like #2 are ridiculous. If you can’t see that, then I feel sorry for you as baseball must be pretty predictable and boring, of course, until all your predictions fall apart. The intelligence in Dave, Derek, Jason, Jeff, and everyone else that contributes or has contributed here is the admission that tools like sims only a small part of what a team like the Mariners needs to look at to be a success. How about zooming out a bit before we become caricatures of ourselves…