THE BOOK--Playing The Percentages In Baseball

Brian Sabean and Bruce Bochy earlier this month talked about how the Giants were still in the playoff race and how they were disappointed that they were so far behind and knew they could play better. I personally think it was just a ploy to try and not scare off any more fans from the beautiful ballpark in San Francisco, and after Sabean’s trade of Durham (for 1 mediocre guy and 1 crappy guy), it’s safe to say that these two (or at least Sabean) are not delusional.
I guess sometimes it’s better to be positive and wildly optimistic than negative and realistic.

I’m an Astros fan, but a realistic Astros fan; when I heard about this trade I almost blew up.  As most knowledgeable Astros fans know, these kind of shenanigans are strongly driven by owner Drayton Mclane’s inability to accept rebuilding or defeat of any kind.  Wade may not be the greatest GM ever, but he’s working within constraints that surely must be intolerable.  Maybe he’s just a masochist.

I’m just glad they didn’t trade Hunter Pence away for Barry Zito.

Anyone want to buy the Astros?  Mr. Cuban?

I don’t think the trade is that big of a deal like people are making it out to be.  It’s more about putting a better team on the field, which is probably why the GM made this trade.  Fans pay money for tickets, why not give them better players.  It’s not like they gave up a young prospect and it’s possible they could get a supplemental pick out of it if they offer Wolf arbitration.  It’s just not worth the breath to bash Wade on this one imho.
vr, Xei

I’m an Astros fan too, and my read is similar to Andrew’s. McLane told Ed Wade to go for it this year, and Wade is doing what he can. That’s the only way I can explain all those off-season signings.

I thought at the start of the season that the idea had a chance to work—not a good chance, but a chance. But the Central has turned out to be much stronger than I expected, and the moves didn’t work out as well as they might have. (Hello, Michael Bourn and your .280 OBP.) Also, hard to win giving up 111 homers in 100 games. Certainly, this result was more likely than a playoff spot.

It will take a crash-and-burn to get the Astros into rebuilding. At least they have only one really bad contract—Carlos Lee’s.

Padre fan here, pointing out that Wolf could enjoy incentives with his contract.

$1.75 mil if he makes ten more starts, plus $250,000 at 190IP or $500,000 at 200 IP according to Cot’s.

I’m not buying the Rockies being out of it at 6 games back.  They may not be a good team, but there is no good team in that division.  Somebody’s still going to win it.

I don’t see the Rockies being 6 back of Arizona as anything different than the Yankees 6 back of the Red Sox.

The AL wildcard is in play for the Yankees. There is no way on God’s green earth that any team out of the NL West is going to win the NL wildcard. That’s where the biggest difference is for the Yankees versus the Rockies.

And do you mean the Yankees from last season? Because right now the Yankees are 3.5 back of the Rays, at least as of this morning.

Rally,
Facts are stubborn things. MGL has the Rockies at 2% to make the playoffs for a concrete reason. They would have to go a minimum of 34-26 to reach the playoffs. And that assumes only 78 wins to get the division, so that’s a best case.

http://www.coolstandings.com/baseball_standings.asp?i=1

Cool Standings has the Rox at 3.5% of making the playoffs, 7 games out, and in 3rd place.

Yanks are 3 1/2 out, in 3rd place, with a 15% chance at the division, and 32% chance of making the playoffs.

If these are outlandish, then the best thing to do is to bet against it.  What does Tradesports.com show?

Rox last traded for 0.9% to win the NL pennant.  With 4 teams to make the NL playoffs, we can multiply by 4 to get a rough approximation.  So, they’ve got say a 3.6% chance of making the playoffs, then 1.8% to get to the LCS, and 0.9% to win the NL pennant.  Seeing that they are a bit worse than the other teams that would make the playoffs, maybe it’s more like 4.5% to make the playoffs, 2.0% to get to the LCS and 0.9% to win the NL pennant.

Again, the best thing for people to do is to come up with their own numbers as to what they would bet, and make sure you get a total of 4.0 teams in the playoffs and 1.0 teams to win the NL pennant.  And anywhere where you differ from the line, bet against it.

By betting, I mean to bet chips or other things that are of non-monetary value.  Unless you live in a jurisdiction where betting is legal of course.

I know the Yanks are closer than 6 games, out, it was just an illustration. Just saying a bad team should have just as much chance catching another bad team from 6-7 down as a good team chasing another good team.

Of course my “playoff odds” numbers are highly predicated on each team’s expected wp for the remainder of the season.  That will differ from one source to another.

I have the Rox as a .480 team and ARI as a .540 team (which might be too generous actually).  I have LA as a .520 team, which is probably around correct.  I have SD as a .480 team also, despite their horrendous record.

I have the Yankees as .530 team, better than TB (.520) and worse than the R Sox (.540), but as Tango says, the gigantic difference between the Yankees chances and any team in the NL West is the wild card. 

I only have the Yankees (as of Wed) with a 15% chance of winning the div, but a 21% chance of winning the WC.  Tha Rox have a 1.3% chance of winning the div and a ZERO chance of winning the WC (0 times out of 10,000 sims)!

BTW, any “playoff odds” simulator that uses a team’s current record as their record for the remainder of the season, is a BAD one.  Even if that is all the info you have, you would still regress that wp to either .500 or some combination of .500 and maybe last year’s or the last couple of years’ wp for that team, if the team has not changed all that much.

With all the good and freely available player projections out there, how tough is it to aggregate all the player projections on a team and figure out their “true” projected wp?

You would be surprised how divergent a team’s actual record can be as compared to what you would get from the individual player projections, even 2/3 of the way through the season.  For example, ATL is way under-performing, FLO is way over-performing, SD is way, way under-performing, CLE is way, way under-performing even with all their injuries, TEX is over-performing, ANA is over-performing, DET is under-performing, and there are a few others.

In fact, if anyone wants a small (legal, of course) wager, I’ll take any of those above teams on their wp hence versus their wp so far, and I’ll spot you 2 wins!

There are only 6 teams out of 30 that never went into the post-season in 10,000 Monte Carlo sims, as of the end of Sunday, the 25th:

SEA
KCA
BAL
WAS
SFN
SDN

Of course if I ran enough seasons, even those teams would eventually win a post-season berth and even a World Series!

Of the 10,000 I run each day, presently, these are the teams in addition to the above teams, that never won a World Series:

CIN
HOU
PIT
CLE
TEX

These are the teams that won the WS the most:

ALA 14.8%
CHN 12.9
BOS 10.9
TBA 8.7
CHA 8.2
MIL 7.9
NYN 7.4
ARI 6.2
NYA 5.8

Another delusional GM, J.P. Riccardi:

“Why would we trade him? We’ve worked too hard to get to this point to break up our team in August.

“If we have a good series here we’re right back in this thing. Now we just have to go out and do it.”

Uh, because you have a 2.8% chance of making the post-season…

It seems to be the dumb GM’s and the dumb teams so far that are delusional about the their chances of making the playoffs.

Am I calling the HOU, COL, and TOR GMs and teams “dumb?” You bet!

Nothing like a 5-game win streak to delude the GM into thinking he has a chance, or a few good starts in a row to make him think that what was previously his trade bait is now untouchable.

If JP fails to trade Burnett and loses him at the end of the year, it will essentially cement his case as a bad GM (if it’s not cemented already.)

The delusional Astros are now only 2 games back of the wild card.  Regardless of what they do in the next 2 weeks, it’s been a great run for them, there is value in having a team contend instead of waving the white flag in June.

The daily playoff odds reports are a relatively new thing, and I suspect they would not hold up to empirical testing.  Seems like there’s a team or two like this every year that seems buried and yet fights their way back to win a playoff spot (like last year’s Rockies) or at least keep things interesting until the final weekend.

The 1 in a thousand chance doesn’t seem to be accurate.

Why would you think that they would not hold up to empirical testing?  I am sure quite that they would.  Not that there would be nearly enough of a sample size to test, but there is absolutely no reason why a decent model would not be accurate.  Heck, a good estimate of a team’s true wp is not even that important.

I doubt that there are a “team or two” every year that is buried that fights their way back.  What, you think there is some kind of magic or fairy dust in the air?  I think your selective memory is getting the best of you.

Without studying the issue, it seems like Rally’s got a point. Over the last 5 or 6 years, it does seem like there’s one team out of 15, left for dead, that comes back to make the playoffs.

If you add up the odds of those 15, it probably comes out to 30%.  So, it is fairly realistic that if you throw a dice, you can get 1 or 2 for 6 straight rolls.

However, if the question is: “what are the chance of a team will be within 2 games of the divisional or WC contender at some point in September”, then out of those 15 bottom feeders, that 30% probably shoots up all the way to 60% or even higher.

So, Rally has a point that it does look like there’s always Cinderella there.  I don’t think it’s selective memory.  MGL is almost certainly correct that it would hold up to statistical testing.

Yes, what I meant was that whatever Rally is remembering is likely well within statistical realms.  Not that there aren’t teams that climb out of oblivion nearly every year.  Of course there are.  If every year, there are 6 teams with a 3% chance of making the playoffs with a month or half a month to go, there is almost a 20% chance per year that one of them will actually make the post-season.  So of course there are going to be Cinderella stories all the time.

Rally’s clear assumption was that the playoff odds reports are wrong because of this.  That I vehemently disagree with simply based on the notion of, “Why would they be wrong?  All you have to do is approximate each team’s true wl percentage, which any idiot can do after 5 and a half months, and then sim out the rest of the season.”

There is no magic fairy dust that is going to make the sim methodology invalid.  I am suprised that someone as smart as Rally would even think that (that because there are Cinderalla teams all the time, the playoff odds reports must be making some kind of error).

http://www.CoolStandings.com does one… uh, cool thing.  In addition to running their sim based on some “smart” way, they also run their sim based on presuming each team is a true .500 team.

(Select dumb on the right hand side of their home page, and click GO.)

http://www.coolstandings.com/welcome.asp

“Another method is to simply assume that any team has a 50-50 shot of beating any other team. You could flip a coin to decide who would win each game. This method isn’t too realistic, but it usually gives Chicago teams a better shot at the division. For lack of a better term, we call this prediction mode “Dumb mode”. “

What is nice here is that a team that is pretty much below .500, and almost certainly a true under .500 team now given a much better shot of making the playoffs.

Unfortunately, it doesn’t look like Cool Standings shows the historical data for the Dumb method, just the “Smart” one.

They show the Astros as under 0.1% as of Aug 26 (66-66), and at 12% on Sept 13 (80-67 record).

They’ve gone on a 12-1 tear.

http://www.baseballprospectus.com/statistics/Astrospsoddspec.php

They have an obvious bug on Sept 1 and 2, probably related to Clay’s computer issue.  (Maybe someone can alert BP on this.) Anyway, Aug 27 they were at under 0.1% and at 11% on Sept 14.

I’m not sure that the expectations should have been any different, unless we believe there is some “reversal of fortune”, such that a team that is really out of it is more likely to go on a 12-1 tear, or what the Rockies did, than a great team that has already clinched.

That would be a fair argument to have, but one that would need to be studied.

"If every year, there are 6 teams with a 3% chance of making the playoffs with a month or half a month to go, there is almost a 20% chance per year that one of them will actually make the post-season.”

I agree with this.  But if you have 10 teams with a 1 in a thousand chance of making the playoffs, which is where the Astros supposedly were, then you shouldn’t see one of those teams make the playoffs more often than once every hundred years or so, right?

I wonder what the 2007 Rockies chances looked like in August of last year.

According to Coolstandings, which despite any criticism I may have, is a great site:

2007 Rockies: They went through a slump in late May, and their playoff chances were down to 1 in 500.

2006: On June 10th their chances were less than 1 in a thousand, buried 11.5 games behind two teams (Detroit and Chicago) playing .600+ ball.  The Twins then got hot and eventually passed both teams.

2005: On June 15th, Oakland was 9.5 games back, with a 27-37 record.  Houston was even worse, 26-38 and 14.5 games back.  Both had playoff odds less than one in a thousand.  Oakland didn’t make the playoffs, but they did catch and briefly pass the Angels before finishing second.  Houston went to the World Series.

So next time I hear that some team that’s 8-10 back with 3-4 months to play is delusional for trying to compete, I’ll remember the words of the Yogi.  It ain’t over till it’s over.

I actually never said someone is delusional for trying to compete.

First I asked the question if a person is delusional or if he thought his fans are.  Even if you take it that I was calling him delusional, I was asking with regards to becoming buyers in the market.

http://www.insidethebook.com/ee/index.php/site/comments/is_paul_depodesta_delusional_or_did_he_think_his_readers_might_be/

He asked his readers what the Padres should do, and 87% said to sell while 7% are delusional enough to want to buy… Obviously “you never know”, as recent Astros and Twins teams have shown.  But, if you’ve got 100$ in your pocket, do you spend that on a rollercoaster ride that may work only 1% of the time, or do you buy some high-flying stocks that may pan out 25%+ of the time?

You compete all the way to the end, and beyond.  Of that, there is no question.

The delusional comment was directed toward being buyers, which is not the most efficient way of behaving.

I don’t think we’ll have any argument that the players on the field should compete to the best of their ability, regardless of their place in the standings.

Where it becomes an issue is if you have players who even at less than 100%, are clearly the best options for their teams.  Do you keep running them out there or have them take a head start on their offseason rehab (Troy Glaus is doing this right now for the Cardinals, to the detriment of my fantasy team).

And as for being buyers, Randy Wolf was a nice pickup for Houston.  If you are longshots, you can’t give up top prospects for an upgrade like Milwaukee did for CC, but if you’re just trading marginal prospects and spending Drayton’s money, I don’t think they were being delusional.  They were just trying to be that one team that comes back from the dead every single season.

I think the argument is getting sidetracked here, which is fine.  The original contention by Rally was that the playoff odds reports would not hold up to empirical testing.  The anecdotal evidence he gives (presumably) to back up that claim is that he thinks there were a lot of teams that came back from oblivion, more so, statistically, then the playoff odds would suggest.

My only contention, whether that is true or not (that more teams have come back from oblivion than a “playoff odds report” would suggest), is that the playoff odds reports are relatively accurate, if only because they are easy to do.

Now, what a team that has a 1 in a 1000 or a 1 in a 100 or 1 in 30 chance of making the post should do, strategy-wise, is another story altogether.

As well, whether you want to consider a 1 in a 1000 or a 1 in a 100 or even a 1 in 30 team “in it” or “out of it” is a matter of taste, like chocolate or strawberry ice cream.  There can be no argument using words like that - only the numbers themselves can be argued.

By, the way Rally, a friend of mine got hit my lightning a few years ago.  I tried to argue that it must have been something else or he was delusional because the chances of him getting hit by lightning were like 1 in 10 million.  He did not appreciate my logic.

I think the arguments are related.  A team that really has less than a one in a thousand chance of making the playoffs, their management should be playing for next year.  You should almost never find examples of such teams actually making the playoffs.

Since in just 4 years of Cool Standings data I can find 3 examples of such teams making the playoffs, and a few others that made it interesting, then I suspect they should not have been considered such longshots in the first place.  Maybe they should have been 1 in a 100 or 1 in 30.

In my opinion it is absolutely worth it for such teams to make reasonable efforts (trade a grade C prospect yes, trade a grade A prospect for a rental player, no) to pursue the chance of a comeback.

It may make business sense (not turn off your fanbase down the stretch) and may give your team a shot at baseball immortality - great comebacks and great collapses make for great baseball literature.

The Astros are, what, still only the fourth-most-likely team to win the Wild Card at this point? (Have to include the Mets in the picture as well because of how tight the East is.) If the Astros don’t make the playoffs, I really fail to see how they’re a commentary on the playoff odds reports at all.

Since in just 4 years of Cool Standings data I can find 3 examples of such teams making the playoffs, and a few others that made it interesting, then I suspect they should not have been considered such longshots in the first place.  Maybe they should have been 1 in a 100 or 1 in 30.

Unless you want to do some kind of analysis that suggests that the coolstandings playoff odds on certain teams or in general were bad - which you are welcome to do of course - I find that fact that you say that there were 3 such examples, less than a convincing argument of anything…

Next year and the years after, if we find 10 teams that have essentially no chance, a 1 in a thousand shot at the playoffs, do you think we’ll have to wait 100 years to see one of them come through with a playoff spot?

I’ll take past observation, that says one of those teams makes it anyway every year or two.