A quick look at division series playoff and luck

by RollingWave on Jul 22, 2009 10:00 AM EDT Comment 14 comments

During the Halladay discussions one of the obvious point that was brought up is how much. or if at all Doc Halladay improves a teams' chances once in the playoff. one of the studies that was meantioned of wanting to get to is the division series.

So I decided to take a quick and dirty look at the division series, since it's not exactly a huge sample. I'm going to quickly see wether.

1. the team with the better regular season winning % had a better chance of winning and how much so

2. wether the team with the better team ERA had a higher winning % and how much so . since the whole "pitching wins playoff" is getting towards mythical status.

So the answer, since 1995, in the division series

Teams with better winning % won : 25 times

Teams with worse winning % won : 29 times!

Teams had same winning % : 2 times

Teams with better team ERA won : 34 times

Teams with worse team ERA won : 21 times

Team had same ERA down to triple digit decimal : 1 time.

So. while the sample size is questionable. the basic conclusion we might draw is.

A. winning percentage does nothing. teams with better winning percentage have no better chance of winning than tams with worse.

B. Team ERA have a positive correlation with winning the division series. but the correlation is far from abosalute. teams with better team ERA won about 63 % of the time.

0 recs | Comment 14 comments

Story-email Email Printer Print

Comments

Nice to have that data. Quick and dirty can still be interesting.

If nothing else, this points out (yet again) how much of a crapshoot the playoffs really are.

Beyond the Boxscore Not a member? Sign up.

by Sky Kalkman on Jul 22, 2009 10:20 AM EDT reply actions 0 recs

One simple way you could expand on this information is to "bucket" the win% and ERA differences.

Maybe separate “close” and “not close”? Sure it would be a bit arbitrary. Maybe pick a few different cutoffs so we can tell if one choice would be cherry picking?

Beyond the Boxscore Not a member? Sign up.

by Sky Kalkman on Jul 22, 2009 10:22 AM EDT up reply actions 0 recs

yeah

a lot of room to tinker with here.

a. the match ups : since this is what people generally look at, the ZOMG TEH ACEZZZZ WINZ PLAYOFFZZZ logic. but this is arbitrary too. since the guys with the better career may not have the better seasons. and who have better careers is subject to debate most of the time anyway.

the underlying stats. using FIP or similar numbers and adjusting for park factor would be better. the Rockies having a 4.50 ERA is not the really worse than the Padres having 4.40 ERA.

c. the close and not close : obviously a team with a 0.1 ERA seperation is not the same as one with a full run. nor is a 100 win team vs a 80 win team the same as a 95 vs 94 . (though this would require a much more thourogh data input to make any sense of )

But really, right off the top of my head we could see some obvious problems. for the more recent years.

2008 : the Phillies have basically one good pitcher and a lot of meh. Tampa have 3-4 very good pitchers . yet the former won decisively.

2007 : the Red Sox were clearly just the better team in every respect anyway.

2006: the most epic case of bad team winning it all in wild card era.

2005 : really? is someone seriously going to attempt to debate that Contreras / Buerhle / Garland are superior pitchers to Clemens / Pettitte / Oswalt? whats worse is that both Clemens and Oswalt threw up stinkers. and the best pitching performance by any starter was ummm… Brandon Backe?

So if we look at the recent 4 WS (granted it’s a cherry picking cutoff) only once did the obvious better pitching team won.

by RollingWave on Jul 22, 2009 11:53 AM EDT up reply actions 0 recs

cherry pick!

even though the Astros had a great rotation, against most other teams, the sox wouldve had the better rotation

by The_Fan on Jul 22, 2009 2:12 PM EDT up reply actions 0 recs

yeah

but the Astros front 3 was really REALLY awesome that year. and career context wise they are clearly better than their White Sox counterpart (actually, all 3 Sox added up probably isn’t even close to Clemens’ worth) with the general babbling on how ace basically make you imprevious to any other deficencies. you would think that the Astros shoulda like.. at least not get swepted.

by RollingWave on Jul 22, 2009 10:51 PM EDT up reply actions 0 recs

CLUTCH!!!

@bs_uf15bosox9be:OverTheMonster-ALLERGEN WARNING:May contain PB.

by bs.uf15bosox9bears23 on Jul 23, 2009 12:00 AM EDT up reply actions 0 recs

I want to see this myself. We could look at something like regular season Pythagorean vs. regular season FIP since 1995 (I want to try out tRA*, but that seems hard to find). I’m looking into it…

Marlin Maniac, a Florida Marlins blog

by SFiercex4 on Jul 23, 2009 1:23 AM EDT reply actions 0 recs

I'd figure

that with a large enough sample. ERA and actual records wouldn’t be too far off from TRA and pythag difference should be fairly similar.

by RollingWave on Jul 23, 2009 6:40 AM EDT up reply actions 0 recs

I did the same, but...

split up the DS series to the ones where the WC was involved and the ones where it wasn’t.

With no wildcard team involved, the team with the better record was 16-10; it was 15-9 if you use a minimum gap of 3 games per 162 games.

With a wild card team involved, the team with the better record was 9-17! With a 3+ game gap, it was 6-13;

Despite having a better regular season record just 23% of the time, the Wild card team has won 58% of their series. Is there some inherent advantage of being the wild card?

by erosen on Jul 23, 2009 9:54 AM EDT reply actions 0 recs

The average division winner...

1995-2008 won 95.1 games; the average wildcard team won 92.7; the average non-playoff team won 76.0.

I looked at the average for the rest of their division [ie, the 3-5 teams they play nearly 50% of their season against] aqnd got:

Non-playoff teams division mates average 82 wins; wildcard teams average 81 wins; division winners average 77 wins.

So, your average wildcard team faced a 5% tougher schedule than the team that lead their division. Adjust up the WC teams by 2 wins and their division winning opponent down 2 wins and the series involving WCs is now 12-12; and the team with the better adjusted winning% and a 3 game gap or greater is 7-8…

That atleast gets us closer to something that makes sense :)

by erosen on Jul 23, 2009 10:32 AM EDT up reply actions 0 recs

I am curious why you choose team ERA. Since playoff teams tend to shorten up their staff going into the post season shouldn’t the top or the rotation guys, top set up ( or 2) and closer FIP (or ERA for old school) be a better indication? Those 6 -7 guys should would be the arms throwing more of the innings and more often in leverage situations, no?

Interesting stuff though and thanks for sharing it.

That’s it! Katie bar the door for this 79 win team is star crossed! And I am loving it.

by daveinexile on Jul 23, 2009 11:48 AM EDT reply actions 0 recs

How about taking the team of interest (say the Phillies), lay out their rotation in a five game series without Halladay and assign a mythical win percentage of each game. Then do the same while adding Halladay to the mix. Then run a simulation of a 5 game series a million or so times, and see what kind of probability of winning the series boost they get from him.

I think anyone who is fairly good at handicapping could easily come up with a pretty accurate number using this methodology.

vr, Xei

by Xeifrank on Jul 23, 2009 12:19 PM EDT reply actions 0 recs

Shouldn't be too hard to without a sim, too.

Just compute runs per game for both teams with/without Halladay and use Pythag for each game to get probability of winning.

Beyond the Boxscore Not a member? Sign up.

by Sky Kalkman on Jul 23, 2009 12:28 PM EDT up reply actions 0 recs

As long as you are really good with combinatorics/math then that’s fine. The hard part is getting the game by game winning percentages. You can do it with handicapping, baseball game simulator, your pythag method and perhaps another or two.

Team A win probabilities:
G1: 59%
G2: 57%
G3: 44%
G4: 45%
G5: 59%

What are the odds that Team A wins this 5 game series?
vr, Xei

by Xeifrank on Jul 23, 2009 2:46 PM EDT up reply actions 0 recs