The Rays and the AL East: 2010 Projected Standings
Now that Andy Hellicksonstine has so successfully taken us through the Rays' WAR projections for 2010, I thought it'd be the logical next step to look at how all the projection systems rate the Rays as an overall team. We've looked at these during the course of the post-season as projections have become available, but now that the season is upon us, we can look at all the projections together and get a better overall view.
|
Win Total |
AL East Finish |
|
|
88 |
3 |
|
|
89 |
3 |
|
|
92 |
3 |
|
|
97 |
2 |
For more information on the differences between projection systems, visit here.
All right, now before anyone freaks out, there are a couple key things to keep in mind when looking at projected standings (or actually, any sort of projections).
First of all, projections shouldn't really be called "projections" per se. Yes, we can look at them as estimates of how a player will perform this season, but a much more accurate way to describe them is as a measure of a team or player's true talent level. In the words of Sean Smith, the inventor of the CHONE projection system:
"I don't have many real surprises though, because all I'm doing is looking at a player's multi-year track record, adjusting for league/park/age, and assuming he'll proceed to age like the majority of players in history have done. Some players will put up seasons in 2010 that will be much better than they've ever done before, and do so past the age of 30. I can pretty much guarantee that CHONE will not tell you who those players are. Projection may not even be the best word for what I'm doing. I'm estimating a player's current talent level."
We all know that there is a lot of random variation and chance that happens in baseball. Some ground balls squeak through for hits, while some line drives get caught. Fly balls that would have been home runs in one park become outs in another. A fielder gets a bad break on a ball and it goes for a double. Pitchers walk three batters in a row but get out of the inning without letting up a run. Lots of crazy, ridiculous stuff happens over the course of a season, which is part of what makes baseball such an amazing sport to follow.
With all that random variation, though, how can we tell how good players and teams actually were last season? We know what the results were because we saw them, but how do we measure the talent? That's a key distinction and it's the question that projection systems answer. If we get rid of all the other stuff, what is a player's true talent?
Of course, in this upcoming season there's going to be a whole lot more of luck and random variation. Players will get injured, players will bust, players will boom. Like Sean Smith said, projections are all but worthless for predicting those sort of seasons. Instead, the projections are giving us a player's 50th percentile numbers - the point where if we simulated the 2010 season 100 times, the player would perform better than those numbers in 50 seasons and worse than those numbers in 50 seasons. It's what we believe a player should perform in a neutral luck environment, based on our current knowledge and research.
And so, when you see the Rays projected for 89 wins or 92 wins, realize that the win total is not a definitive answer. It's not saying that the Rays are definitely going to win 92 games and no more. Instead, it's saying that it's likely that the Rays will win 92 games, but through random luck and variation alone, this same team could win anywhere from 75 games to 105. Don't believe me? Look at the range of win totals possible for the St. Louis Cardinals, another team projected to win around 90 games:
|
Win Total |
Probability |
|
<=65 |
0.00% |
|
66-70 |
0.00% |
|
71-75 |
1.00% |
|
76-80 |
5.40% |
|
81-85 |
16.50% |
|
86-90 |
29.60% |
|
91-95 |
28.00% |
|
96-100 |
14.40% |
|
101-105 |
4.40% |
|
>=106 |
0.60% |
While it's most likely the Cardinals will win around 90 games, it's still possible (although highly unlikely) for them to finish with a record below .500. Crazy, huh? This is a great thing, though, because it gives the Rays hope. While it may seem disappointing that we're projected for a third place finish, we're so close in all those projections to the Yankees and Red Sox, anything can happen! Our front-office has got us a team that can compete with the Yankees and Red Sox, so now we're just going to have to see where the luck falls this season. And thankfully, luck is one thing the Yankees and Red Sox can't buy.
12 comments
|
1 recs |
Do you like this story?
Comments
I went ahead and looked at the Yankees and Red Sox in the same manner as I did the Rays yesterday
I’ll drop that knowledge a little later, should sandwich around this nicely. I agree that it’s important to stress the 50th percentile idea behind this. It’s a bigger deal that we are within a few games than what those numbers are. It should be the most exciting Rays season yet.
"It's good to have a little cushion. But it's not going to be easy."
by Andy Hellicksonstine on Mar 16, 2010 1:28 PM EDT reply actions
These projections are all versions of Monte Carlo simulations.
It would be interesting to know what the bottom 20% and top 20% (or even a 1 Standard Deviation) of the simulations came out to in wins.
Here's to swimmin with bowlegged women.
According to my projections...
I have the Rays as 2010 AL Champs with a 99-63 record and winning the WS to NYY in 5 games. Boston wins 95 with 66 loses(one game rained out will not even be necessary to play). I will reveal my formula by December 1st, 8:00 pm in this forum -Please don’t ask before then. Thanks
"A chain is no stronger than its weakest link."
Two problems with this prediction:
1. It doesn’t make any sense to wait until the end of the year, let alone the season
2. “This forum”, specifically this thread, will have long been locked.
"Sure, because of the "cold weather" and rain." More bait and switch tactics by the New York owners of this team." --NikoHoullis, the lead blogger at Buc'em on racial and anti-semitic insensitivity.
The first problem with this is
the Rays playing the Yankees in the World Series.
From there, I guess anything is possible.
by nyyfaninlaaland on Mar 17, 2010 2:24 AM EDT up reply actions
I knew it was not going to last long...
I’ll try again without using the “floating rearrangement” of the divisions.
"A chain is no stronger than its weakest link."
by I Believe-D on Mar 17, 2010 10:23 AM EDT up reply actions
Calling projections a look at "true talent level"
isn’t exactly accurate either.
Do players talents shift so dramatically from 1 season to the next? Projections do change as a player puts up a broadening and evolving set of performances. The small sample data sets of young players mean higher potential deviation. And aging curves, while useful for the broad midsection of players, might not apply in quite the same manner for top performers. The upside is sabermetrics remain a work in progress – be kind of boring if nothing was left to do, no? And still it all comes down to the individual deviations from the mean. Few players are as consistent as Henry Aaron.
Was Ben Zobrist’s 2008 projection his “true talent” level? At the time Zobrist was regarded as a virtually a replacement level player.
Was Burrell’s 2009 projection his true talent level? Is it also so for 2010? Burrell does have a rather substantial data set to go by, while moving along some aging curve trajectory. His seems a bit precipitous at present, but even that could change.
Projections are statistical analyses and estimates based on the available usable and relevant data. The variations in the possible outcomes – and kudos to Steve for emphasizing this as he does here – are the intersection of statistical compilation and individual case variation.
Meaning projections are basically very educated guesstimates at this point in time – with emphasis on the very educated part, not the guesstimate part. Nothing more. Not meaningless, but certainly not set in stone.
by nyyfaninlaaland on Mar 17, 2010 2:45 AM EDT reply actions
Good points, but there are one or two little critiques
I think you’re short-changing projections a bit. I tend to look upon player projections the same way I do the projections of team win totals. If we’re predicted to win 90 games but only end up winning 80, where should we estimate our team’s actual talent level was? Closer to 80 or closer to 90? You need to look at the data to determine why you under-performed, but then you can find if the change was a result of luck and variation or the result of a change in talent level. The Rays last season, for instance, under-performed their win total by a large amount, but we’re not worried about being a .500 team again this season. That’s because the projection is a better estimate of our true talent level than our actual performance was last season.
Of course, there are going to be the players that break the mold. Cherry-picking the two players last year that performed much different than their projections doesn’t mean the whole system isn’t overall pretty accurate. Yes, the projections couldn’t predict that Zo would explode last year or that Burrell would be injured and age quickly. But based upon the data we have, going into the season their projections were the best estimates we had of their talent levels. With the new data from last season, now both the evaluations have changed. Their changes in results were a combination of changes in talent level and luck and variation, so we had to shift their projections this season.
It depends how much faith you have in the projections, but I tend to look at them as the best measure we have available to measure a player’s true talent level – very educated estimates, although I know that’s just a semantic difference from what you said. If their results differ widely from the projections, then we need to decide how much of it was a change in talent level and how much was luck and variation. Of course it’s not going to be 100% accurate, but more often than not the projections are closer to measuring a player’s true talent than their one season’s results are.
I love Casey Fossum. Now try and take me seriously.
by Steve Slowinski on Mar 17, 2010 9:48 AM EDT up reply actions
MLB The Show 10
I’ve simulated the upcomming season with the probable 25 players on my PSP with MLB the Show 10. Joyce platoon with Kapler in RF, Zobrist on 2nd base, S-Rob on the bench. Davis as the 5th starter., Sonny in the bullpen as a long reliever. The result was 88 wins and third behind Boston and NY. For what it’;s worth….
Now do that 1000 times
As for your point above, nyy, the projections work pretty well when looking at an entire league, when the variance between who’s above and who’s below washes out. It’s not meant to be a perfect predictor, just a signpost.
"It's good to have a little cushion. But it's not going to be easy."
by Andy Hellicksonstine on Mar 17, 2010 10:29 AM EDT up reply actions





















