16 replies to this topic

[DocPeterWimsey]

I've mentioned this before, but I thought that people might like to see it in a little more detail. The post-season is a bit of a crapshoot. However, there is a general tendency for teams who played well in the last 4 weeks of the season ("September") to do well in the playoffs. In particular, relative performance is a good predictor of who will win the first round.

Here are the September run differentials from the 44 teams that made post-season after 2000. The bars shaded in black are those of teams that lost in the first round:

Perhaps unsurprisingly, most playoff teams played well in September. This is not because "that's when it counts" so much as good teams usually play well. What really stands out is that the losers (in black) are clustered towards the left. In particular, teams that fell apart in September (say, the 2008 Cubs, Brewers and ChiSox) and post negative run-differentials tend to lose.

Of course, the game is not played in a vacuum: and if you have a hot team with a +60 run differential (RD) playing a hotter team with a +80 RD, well, then you might think that the +80 team has the better shot. (This has not happened in the last 11 years, but you get the idea!)

So, here's what we get when we look at the difference in September Run Differentials between the winner and the loser.

I've added the 3 Cubs series because, well, this is a Cubs site! The Cubs do not stand out in the slightest: the 2003, 2007 and 2008 LDS all were won by the team with the better September.

This is the only breakdown I can find that predicts significantly more than 50% of the winners. Basically, half the time the team with the most wins or greatest rund-differential for the entire season takes the series. Slightly more than half the time, the team with the best actual won-loss record wins: but the run differential (which predicts W-L with some error) does a much better job.


So, as a rule of thumb, bet on the team that had a better run-differential in September. Such teams win about 3 times in 4. Last year was a good example: the Cards, Tigers and Rangers won but the DBacks lost. Now, you say, what good is this if you cannot tell me which 3? Well, tough: probability does not work that way!


I will add this caveat: the new 1-game WC playoff might very well alter this by making a "hot" September team use up their Ace and thus getting him for only one of the LDS games. At this point, I'm not sure that the managers have decided how they will handle this game: it probably will take a few years before any tactic emerges as superior.


Any way, food for thought....

[fromthemitten]

just want to point out the first year I moved to NYC for college the Yankees won the World Series and now I moved to Maryland and the Orioles and Nats are in the playoffs.... somebody should pay my rent to live in Chicago next year.

just sayin'

[OCCubFan]

Very interesting and illuminating. Am I correct that even the September run diffferential is not a good predictor of success after the first round of playoffs?

[OCCubFan]

Where can I find the run differentials for this September?

[DocPeterWimsey]

Very interesting and illuminating. Am I correct that even the September run diffferential is not a good predictor of success after the first round of playoffs?

It seems to be a poor predictor so far. However, the sample size also is half the size, so there has not been as big an opportunity for differences to emerge. Moreover, you often are getting LCS between two hot teams: and it's quite possible that after a certain point (say, +30 for September) it's really all the same.

It's possible that what this really reflects is the tendency to cull teams playing poorly, and they do not make the LDS often enough to create the same effect. For example, the three teams with negative Sept. run-differentials who went on to the LDS all lost there: but that's not enough to really influence a pattern with 20+ examples.

Where can I find the run differentials for this September?

The easiest way to get it is to download the standings data for the end of the season from ESPN or from MLB and the data on August 31s (or 4 weeks before the end; it won't affect things much), and just subtract the August 31 RD from the final RD. That is all I did!

[DocPeterWimsey]

By the way, here are the run differentials since 3 September for the 10 teams in question:

Tigers 19

A's 20

Yankees 50

Orioles 38

Rangers -20

Giants 26

Reds -3

Nats 23

Braves 18

Cards 20

The Reds are in pretty bad shape: just as in 2010, they played poorly in September given run differential (75 scored, 78 allowed). For what it's worth, the 78 runs allowed was pretty good: only the Nats allowed fewer. And that tells us a lot about how bad the Reds hitting is right now. As they are playing the +26 Giants (128 scored, 102 allowed), the Giants should be heavily favored.

The Nats played well, but so did the Braves and Cards. Still, the Nats did a little better, so all else being equal (and it's not: the Braves-Cards winner will be a little depleted), expect the Nats to win.

In the AL, the Yankees had by far the best September. What's a little sad is that the O's had the 2nd best September of the post-season: if they get past the Rangers (who were awful), then the Yanks will have the upper hand. Moreover, with this one game format, it probably is only a little better than a coin-flip that the O's reach the LDS! (That sucks, in my unhumble opinion....)

The A's and TIgers basically were the same.

In the AL, the Yanks were on fire this September. (In 2010 and 2011, they had very poor Septembers.)

[OCCubFan]

Thanks. This is great. We will, of course, revisit this topic after each round of the playoffs.

[Tommy]

Moreover, with this one game format, it probably is only a little better than a coin-flip that the O's reach the LDS! (That sucks, in my unhumble opinion...)

I agree. Love that they added a wildcard team, but HATE that it's a one game playoff. After 162 game season, I think it's crazy to decide things with a one game playoff. This system doesn't really favor the better team, but rather the team with the better ace.

I'm sure the length of the season played into things, and starting earlier in the year is most likely out of the question, but I still think there's got to be a solution. Maybe shorten the season and get the post-season back into October? I dunno, but April and November baseball is pretty rough in some cities when the temps start to drop (and I have attended a game at Wrigley in early April, brrrrr).

Not sure how you do it, and I wouldn't be crazy about a shorter season as I love my baseball. I just really, really dislike a one game playoff for a league that plays 162 regular season games. It just seems unreasonable to me.

[DocPeterWimsey]

I just really, really dislike a one game playoff for a league that plays 162 regular season games. It just seems unreasonable to me.

It's a little off-topic, but not entirely, as it could affect the tendency for "hot" WC teams to beat "cold" division winners: what is the best tactic here? On one hand, you lose and you go home. That says, go with your best pitcher if he's available. But, on the other hand, what is the point of winning this one game if you seriously hurt your chances of advancing to the LCS? In the end, you are (to paraphrase Doc Gooden from 1988) just like the Cubs: watching the rest of post-season at home.

I realize that this really is an old question: once you get to post-season, the point is to win the World Series, and you can take the view that this means do anything and everything to get there vs. trying to set it up so that you are in the best shape once you do get there. This just exacerbates the quandary.

I honestly have no idea if there is a right answer or how this is going to affect LDS outcomes. The only thing I can think is that it will favor teams with good 4th starters over other teams.

[OCCubFan]

I looked at the graph of Winners' September RD minus Losers' RD and tried to see if I could torture the data and improve upon saying the team with a net RD advantage wins three-fourths of the time. Of the 8 series in which the net RD advantage was greater than 40, all 8 with the superior RD won. Teams with a net September RD advantage of less than 40 runs won 24 of 35 series---call it two-thirds of the time.
If the net RD were zero, then one would expect the probability of winning a series to be 0.5. I then conjectured that the probability of winning a series varies smoothly from 0.5 at a 0 net RD to something close to 1.0 at a net RD of 40----keeping in mind that we are taking a small sample and turning it into a minuscule sample. But that did not fit the data. Actually, teams with a smaller net RD advantage did better on average than teams with a larger, but less than 40, RD advantage. Because this is counterintuitive, it is probably due to the tiny sample size.
Thus, my conclusion is that having a net September RD advantage of 40 or more gives a team very good chances of winning, but advantages of less than 40 runs are less predictive, even though teams with such an advantage have won two-thirds of the series in the past. For this year's playoff teams, it implies that the Rangers' chances of beating any other AL team are pretty slim.

[OCCubFan]

With regard to the tactical issue of whether to use one's ace pitcher in the one-game playoff, I come down hard on the side of "yes". You can win a 7-game series even if your ace loses one or even both of his starts---hard, but feasible. However, there is no such margin for a 1-game "series". That game is the highest leverage game of the potential 8 games; therefore, go with your best.

[SirCub]

With regard to the tactical issue of whether to use one's ace pitcher in the one-game playoff, I come down hard on the side of "yes". You can win a 7-game series even if your ace loses one or even both of his starts---hard, but feasible. However, there is no such margin for a 1-game "series". That game is the highest leverage game of the potential 8 games; therefore, go with your best.

I agree completely. Only time that I think you don't roll out your best pitcher, is when your best is kind of 'meh' and you have a deep enough bullpen to play the platoon match ups virtually all game. Most pitchers decrease in effectiveness their second time through the order, so why not switch it up if its potentially your last game of the season?

[DocPeterWimsey]

Interestingly, neither the Cards nor the O's went with the tactic of "win at all costs now!" yet both won. However, the schedule and competition might have decided things: both teams were in the midst of races (yes, the O's had clinched post-season, but they were trying to catch the Yanks: and it really makes a difference now).

At this point, the single best bets are:
Yanks over O's
A's over Tigers
Giants over Reds
Nats over Cards

Most probably, 3 of those 4 will be correct! That written, only Giants over Reds has the making of a lop-sided series: the Reds are pretty awful right now whereas the Giants are playing well. All of the other teams are playing good baseball right now.

[King Jeff]

That written, only Giants over Reds has the making of a lop-sided series: the Reds are pretty awful right now whereas the Giants are playing well.

Now watch the Reds go out and sweep the Giants.

[DocPeterWimsey]

That written, only Giants over Reds has the making of a lop-sided series: the Reds are pretty awful right now whereas the Giants are playing well.

Now watch the Reds go out and sweep the Giants.

It's improbable, but not as improbable as the '06 Mets sweeping the '06 Dodgers. Similarly, the '06 Cards beat the '06 Padres. Those were two of the biggest upsets given September records, and both happened in the same week! I don't recall the details of the Cards-Pads series, but I did watch the Mets and Dodgers. Two of the games were decided on crazy flukey plays (including blown calls), with the games not dissimilar to the Braves-Cards game last night.


Incidentally, here is the breakdown on the number of series (out of 4) where the "better September" team won over the last 11 years. The number in parentheses gives the expected number of times we should have seen that given an expectation of 3 of 4:

0: 1 (0.0)
1: 0 (0.5)
2: 1 (2.3)
3: 6 (4.6)
4: 3 (3.5)

So, anything can happen: but not everything is equally probable.