Jump to content

PECOTA takes on standings


DrungoHazewood

Recommended Posts

Now you are predicting in retrospect, eh? I thought your retrospective prediction system would have been much better than being off by just five teams. :)

Now try predicting the future. It's a bit harder.

We will see if 99% of the folks here can predict the Orioles standings within one place when we post our prediction poll. We won't even have that high level of agreement, so there will be no chance that 99% of the people will be right.

99% is admittingly high because people will predict with their hearts, not minds.

But if most people project them to finish 4 or 5, chances are they will be within 1 spot of where they finish.

And you know what, those people don't need a project system to do that. Its not hard to predict standings and come within one spot.

It is easy to identify the best and worst teams in each division. After that, you just need to have a few things go right and boom, you are a walking, talking PECOTA system.

Link to comment
Share on other sites

  • Replies 119
  • Created
  • Last Reply
99% is admittingly high because people will predict with their hearts, not minds.

But if most people project them to finish 4 or 5, chances are they will be within 1 spot of where they finish.

And you know what, those people don't need a project system to do that. Its not hard to predict standings and come within one spot.

It is easy to identify the best and worst teams in each division. After that, you just need to have a few things go right and boom, you are a walking, talking PECOTA system.

Yeah, I think Cindy's predictions would probably skew the study.

;)

Link to comment
Share on other sites

Before we refer to PECOTA as psychic I'd prefer to see how they did in the past also. Remember a blind hog will find an acorn every now and then.

PECOTA is still in its infanacy. It went through really dramatic changes prior to 2006 and some decently significant changes prior to this year.

Link to comment
Share on other sites

Wins and saves for individuals are irrelevant. I think if BP wasn't trying to sell books they wouldn't even list them because they're so subject to random variations.

PECOTA was originally a fantasy baseball application and its still is a large portion of their audience.

Link to comment
Share on other sites

I don't quite understand the Roberts projection, though. His top comps are mid-career Ray Durham, Don Buford, Bill Doran, Bobby Avila, Chuck Knoblauch... I think a few of these guys had off years right after their Roberts comparison season but then bounced back. So maybe this is a fluky rating.

The weighted mean is still the meat and potatoes. Roberts blew in 04 and was not super great in 06 and will be 30 at the end of the year.

The biggest issue with PECOTA or any other projection system is injuries. For example, it has no idea that Matsui suffered a freak wrist injury and should be healthy and ready to go this year. Likewise, it doesn't know about Roberts arm injury.

All it knows is he had a bad year, a huge year, and then declined again. The system thinks he is on a huge physical decline, when arguably its just a recovery from injury issue.

THT loves the Os a lot more:

Team           Win     Loss    % Chance of winningNew York       95      67      53%Boston         91      71      32%Baltimore      82      80      8%Toronto        81      81      6%Tampa Bay      68      94      0%

link

Link to comment
Share on other sites

One thing I'm taking from all of this is I'm extremely concerned about is our offense. We are not looking good in ST and I don't want to here this "it's only ST" BS. I'm seeing the same things I saw in 2006 regular season. Can't hit lefties and long streaks of not scoring many runs.

What's ironic is we are signing Rays castoffs to long term contracts in hopes that it will make us considerably better than last season.

Link to comment
Share on other sites

No, but PECOTA also can't factor in highly relevant information in some cases. I think any unbiased human observer would say that a .729 OPS projection for Roberts is very likely to be too low.

But there is a lot of information that a human perspective leaves out. PECOTA has short commings, as Nate's article criticizing PECOTA's projections points out, but so do all of us. Everybody seems to have their panties all bunched up but the reality is that is it just one tool to use. No one should follow it dogmatically, even Nate recognizes this.

Link to comment
Share on other sites

This is bs...Most people felt we would finish 3rd or 4th last year. That is exactly where they ended up.

Pickign the standings is not that hard.

I agree. But lets stop screwing around off topic and address the issue at hand, which is not predicting where a team will finish but what a team's record will be, because that is what Nate is doing.

Link to comment
Share on other sites

But there is a lot of information that a human perspective leaves out. PECOTA has short commings, as Nate's article criticizing PECOTA's projections points out, but so do all of us. Everybody seems to have their panties all bunched up but the reality is that is it just one tool to use. No one should follow it dogmatically, even Nate recognizes this.

PECOTA is historical, quantifiable evidence applied to a future prediction almost like a credit score.

So with that, I guess we can think of it as a baseline that we as fans can use in our own predictions. Knowing what we know about Roberts injury and Mazzone's influence we can predict things a slightly different way. Mazzone's influence though is hard to quantify like many other human factors.

Link to comment
Share on other sites

Its worth pointing out that PECOTA didn't predict these standings.

PECOTA predicts a hitters or pitchers statistical line.

BP than makes playing time estimates and plugs the PECOTA lines into these playing times estimates, then pythagapat's wins and losses and adjusts for strength of schedule. As Nate has written:

We take the raw output from the individual PECOTA projections, put them into depth charts, translate those depth charts into wins and losses, and adjust those wins and losses for each team’s schedule. This schedule adjustment is particularly important this year because of the emerging disparity between the leagues--the average AL club is now several wins stronger than the average NL club.

link

Link to comment
Share on other sites

PECOTA is historical, quantifiable evidence applied to a future prediction almost like a credit score.

So with that, I guess we can think of it as a baseline that we as fans can use in our own predictions. Knowing what we know about Roberts injury and Mazzone's influence we can predict things a slightly different way. Mazzone's influence though is hard to quantify like many other human factors.

Its like any other fantasy baseball projection system. Run the numbers, sprinkle with your knowledge and common sense and mix throughly.

PECOTA isn't perfect but its a damn good resource. You'd be better off relying completely on it than completely not on any system, but the best solution is to use it in combination with your judgments.

This is an issue Tango is trying to address, the whole man v. machine in making predictions. If this interests you I'd urge you to reference this post and fill out your projections for the 2007 Os.

Link to comment
Share on other sites

I think Tejada was pretty bad last spring and started hot.
Yeah, after the trade "demand" of last off season, I remember pretty well that he had a really terrible spring and left a lot of us concerned. Then he goes to hit for the highest average of his career :P
Link to comment
Share on other sites

The weighted mean is still the meat and potatoes. Roberts blew in 04 and was not super great in 06 and will be 30 at the end of the year.

To expand on this idea, look up BR's PECOTA Card and scroll down to valuation, and look at BRs OWarp for 07 (2.4), 08 (2.5) and 09 (2.2).

Why does PECOTA think 08 will be a better offensive year than 07, even though BR will be a year further on the wrong side of 28? Because BRs crappy 04 will be gone, and coming into 08 his past 3 seasons will be the monster 05, decent 06 and whatever he does in 07.

The Player comps have very little to do with the projections, regardless of what they say in the BP glossary. That's pretty much weighted mean. The player comps but are instead primarily used for future valuations, beta (i.e. the degree of certainty the program has in the prediction), the upside breakout scores, the attrition rate, and the player percentiles (which implicitly lift the weighted mean up or down to form the final forecast).

This is why you see the arrows trending up or down in the set of comparable players at the bottom of each PECOTA card. The comparable players "guide" the future forecasts and "soften" or "strengthen" the weighted mean projections.

Link to comment
Share on other sites

The weighted mean is still the meat and potatoes. Roberts blew in 04 and was not super great in 06 and will be 30 at the end of the year.

The biggest issue with PECOTA or any other projection system is injuries. For example, it has no idea that Matsui suffered a freak wrist injury and should be healthy and ready to go this year. Likewise, it doesn't know about Roberts arm injury.

All it knows is he had a bad year, a huge year, and then declined again. The system thinks he is on a huge physical decline, when arguably its just a recovery from injury issue.

THT loves the Os a lot more:

Team           Win     Loss    % Chance of winningNew York       95      67      53%Boston         91      71      32%Baltimore      82      80      8%Toronto        81      81      6%Tampa Bay      68      94      0%

link

I think the failure to factor in injuries is huge. And I think that it is an area where, even if they tried to factor it in, human judgment could probably do a better job.

The Hardball Times' projections look a lot better to me. Not just because of my Oriole bias, either. I agree with them that New York is slightly down, that Boston is up, that Toronto is down and that Tampa is up but not by nearly as much as PECOTA would suggest. If I were to quibble, I'd probably take 2 wins away from New York and give them to Toronto.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...