Iowa State physicist writes a better formula to predict baseball success

He claims consistently scoring teams have a better record as compared with sporadically scoring teams. (with both sharing the same runs per game.)

Could a next movement be to study a players std dev of a stat over time period of weeks or months? Would teams pay for consistency over streaky players?

about 2 years ago tcon125 43 comments 2 recs |

Story-email Email Printer Print

Comments

I think so.

I remember Jeff King being almost useless because he was so streaky. It was maddening because you could almost set your watch to his cold spells. Why they didn’t rest him more in this period is a mystery to me.

by Dadunca on Feb 28, 2010 8:11 PM EST reply actions

I don't think this has any application at the individual player level

its mostly about team run scoring consistency, and the way to insure that is to have a patient team

OBP is much more stable over periods than BA

by Freneau on Feb 28, 2010 8:21 PM EST reply actions

Isn't he saying that a higher Slug begets greater consistency,

not OBP or BA? Also, will someone please alert Joe Morgan (and Ken Tremendous)?

by stuckinstl12 on Feb 28, 2010 8:39 PM EST up reply actions

Find my error:

Wisnant says that adding .080 SLG while holding RPG constant should add one win per year (because SLG correlates with narrower runs distributions)…

…but, and correct me if I’m wrong here, the only way to increase SLG while holding RPG constant is to decrease OBP…

So, now OBP is the enemy? Huh?

And, to make matters worse, we all agree that one point of OBP is worth more than one point of SLG, right? So, it would seem that higher SLG at the expense of OBP ought actually to decrease RPG… I have to be missing something obvious, right?

by kcemigre on Feb 28, 2010 8:41 PM EST reply actions

I just really don't think there are actually any implications to this study

other than, if you slug more you have a lower chance of ever being shut out, and you cannot win a game scoring 0 runs

by Freneau on Feb 28, 2010 8:50 PM EST up reply actions

but the thing is

you can’t start out with your roster and say, “ok, we’re scoring 700 runs this year: how do we want to do it? all power? all batting average? all steals?”

you have to try to score as many runs as possible, period. and the best way to do that is based on OBP, because a better overall roster that scores 720 is going to be better than a consistent 700 run team

by Freneau on Feb 28, 2010 8:51 PM EST up reply actions

What if

a roster that scores 715 has less RPG variance? Isn’t he saying that team would have more wins than the team with 720 runs but more variance?

by stuckinstl12 on Feb 28, 2010 8:55 PM EST up reply actions

yes, he is

and I think he’s right…..but isn’t it kind of obvious? In other words, did we really need a guy with a PhD in Physics to tell us this?

by Crooow on Feb 28, 2010 8:56 PM EST up reply actions

obvious because

extra base hits (especially HRs) are more certain to drive in runs? Thus a team with consistent power will have more consistent RPG? I am Joe Morgan.

by stuckinstl12 on Feb 28, 2010 8:59 PM EST up reply actions

there was an interesting discussion at BP about a month ago

would it be better to have a pitcher that allows 4 ER every game, or a guy that gets his ERA by being all over the place (sorta like Hochevar)

there isn’t necessarily a clear anaswer

by Freneau on Feb 28, 2010 9:01 PM EST up reply actions

if on that day that you get Hochevar'ed

you score more runs.

I would guess that is consistency in RPG is better, also consistency in your pitchers (however you want to measure it by—-I'm guessing Wins) would be worth a few more wins a year as well.

- .... .- - .----. ... / .-- .... .- - / ... .... . / ... .- .. -..

by MarioVanPeebles Republic of China on Mar 1, 2010 1:23 PM EST up reply actions

yea, if that team ALWAYS scored exactly the same number of runs, it probably be better. even the royals will randomly have some 10+ run games with mostly “wasted” runs

by Freneau on Feb 28, 2010 8:59 PM EST up reply actions

right you just want to get rid of those pesky shut outs

- .... .- - .----. ... / .-- .... .- - / ... .... . / ... .- .. -..

by MarioVanPeebles Republic of China on Mar 1, 2010 1:24 PM EST up reply actions

i think the way to look at this

let’s a team with A OBP% and B SLG% will be expected to score C runs over the course of the season which will in turn cause the team to have X winning percentage. what this study tells us is that how C runs turns into X winning percentage is dependent on A and B. we’d already known that a point of OBP% will be more important for scoring runs than a point of SLG%, but what this tells us is that it matters HOW teams score runs. the take-away should be that a point of SLG% may actually be more valuable in generating wins than previously thought, even tho that value does not entirely show up in runs scored. a stat like WAR could easily take this into account, and if other studies reach similar conclusions, it probably should.

by kcdc1 on Feb 28, 2010 9:14 PM EST up reply actions

I think you're imputing a direct

causation between SLG+OBP and RPG. While I think there is an expected RPG for any combination of SLG+OBP, Wisnant is using actual RPG which brings in some randomness. So an increase in SLG does not necessarily mean there will be a corresponding decrease in OBP when holding actual RPG constant.

That’s my guess.

by stuckinstl12 on Feb 28, 2010 8:52 PM EST up reply actions

I think that's it or close to it

- .... .- - .----. ... / .-- .... .- - / ... .... . / ... .- .. -..

by MarioVanPeebles Republic of China on Mar 1, 2010 1:25 PM EST up reply actions

Sal Baxamusa actually had a whole series on this a while back at THT

as he points out in the commets… I don’t know enough to say who is right, although this does look like a case of another academic trying to get into sabermetrics but (understandably, perhaps) missing detailed previous work on the same subject.

http://www.hardballtimes.com/main/article/feast-or-famine-first-draft/
http://www.hardballtimes.com/main/article/avoiding-the-famine/
http://www.hardballtimes.com/main/article/consistency-is-key/
http://www.hardballtimes.com/main/article/consistency-is-key-part-two/
http://www.hardballtimes.com/main/article/consistency-is-inconsistent/

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Feb 28, 2010 9:31 PM EST reply actions

skimmed those articles

but didn’t read in detail. i didn’t see where the two authors disagreed tho. both seemed to agree that slg% is the way to go if you want consistent runs.

by kcdc1 on Feb 28, 2010 9:54 PM EST up reply actions

I don't know, either but as far as the usefulness of the information for team, from "part two:"

So, I’m sorry to say that I don’t have a definitive, quantitative answer, but suffice to say that the way a team distributes the way it scores and allows runs is fundamental to why teams over- or underperform their Pythagorean record. Given that, can teams control their run distributions by intelligent roster construction? I am fairly certain that the answer is no. The correlation between “consistency,” as defined here, and the typical metrics (HR/PA, BB/PA, AVG, ISO, OBP, SLG, etc.) is weak to non-existent. (There may be something to how individual talent is distributed on a team – like a single pitcher or hitter that is much better than his teammates – but I haven’t checked that yet.)

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Feb 28, 2010 10:29 PM EST up reply actions

needs better review

I don’t think there is any reason that academics shouldn’t get involved in sabermetrics (Sal Baxamusa is also an academic). Whisnant should have cited Baxamusa’s work, and the paper shouldn’t have gotten through peer review (assuming this is a peer reviewed conference) without it. But, such things happen. It could very well be that both of them are “right” and merely shedding light on the same problem using two different kinds of flashlights. Perhaps now that Whisnant is aware of Baxamusa’s work, he’ll be better able to put his work in the proper context for the future.

by Trey Hillman's Chin on Mar 1, 2010 11:40 AM EST up reply actions

I'm not saying they shouldn't

or making any accusations. My point is that, as Whisnant himself admits, he wasn’t familiar with all of Baxamusa’s work (just part of it). I think his work would have benefited from that, that’s all. tom tango has a number of things on this as well that have been available for a while.

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Mar 1, 2010 12:46 PM EST up reply actions

I agree

I hope that being better aware of the state of the art allows Whisnant to improve his work.

I also seem to remember there being something about this in Baseball Between the Numbers. I think they simulated games using two teams, one of which had a high probability of scoring a small number of runs each inning, and the other team was a many-or-none type of team. They were both set up to score the same number of runs per season on average. Unfortunately, I don’t remember the result.

I have often thought of how the worlds of sabermetrics and academia could be brought together a little better. It’s a difficult fit.

by Trey Hillman's Chin on Mar 1, 2010 12:59 PM EST up reply actions

It would help if certain academics who do engage baseball

and I’m not thinking of Whisnant here, would actually read sabermetrics without being dismissive, particularly when the sabemetricians do research that demolish to render trivial the “findings” of said academics

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Mar 1, 2010 1:13 PM EST up reply actions

this happens in the academic world too, and you have a recourse

The nice thing about math and science is that there really is no need to respect authoritative sources. The work stands on its own. If some academic is getting published for trivial findings, whoever is able to demolish the work should. If you want it to be accepted in the academic circles these people are running in, you can submit a paper to the same journal, and there should be no question.

by Trey Hillman's Chin on Mar 1, 2010 1:24 PM EST up reply actions

I actually understand how the peer review process works

although in my own field, there’s a much higher BS factor

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Mar 1, 2010 1:34 PM EST up reply actions

BS because of subjectivity or what?

by Trey Hillman's Chin on Mar 1, 2010 1:37 PM EST up reply actions

well, I'd say "subjectivity"

since “subjective” and “objective” mean different things in different contexts

but, yeah, assuming you mean what I think you mean, yes. The lack of testable evidence. I don’t think that invalidates the field, it just means, in our social context, that the usual academic power-play/fashion BS gets inflated in significance by about 500%.

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Mar 1, 2010 1:56 PM EST up reply actions

Another appearent reason that OPS works better then it should.

Go Royals!

by BabyBlues on Feb 28, 2010 10:03 PM EST reply actions

"works better than it should"

what do you mean by that? Just curious.

by Crooow on Feb 28, 2010 10:31 PM EST up reply actions

Because its made by combining two otherwise unrelated stats.

I refuse to set up a signature....DAMMIT

by RoyalPug on Feb 28, 2010 11:48 PM EST up reply actions

OPS is used to predict how effective a team will be at scoring runs

but it weights power too heavily and doesn’t weight on-base percentage heavily enough. this study suggests that a team with more power may perform better than would be expected based on its runs scored. following the study’s results, when predicting wins, one should weight power against OBP% more heavily than one would when predicting runs scored. essentially, this counter-balances the weighting flaw in OPS to some extent and means that it might be better than expected at predicting team winning percentage, if not runs scored.

by kcdc1 on Mar 1, 2010 12:51 AM EST up reply actions

okay

understood. thanks. Pretty cool.

by Crooow on Mar 1, 2010 11:28 AM EST up reply actions

this seems fairly obvious to me

Anything you do consistently will be better than having ups and downs. You will have a clearer picture of things (and therefore can respond easier to weaknesses.) I’d also imagine that streaky teams tend to have more losses when they are bad than wins when they are good. Why? because even when they’re good the other team might be BETTER that day. When they are bad it’s hard for a team to be worse. Given that extreme layman’s version, this seems obvious.

Obviously, you are not a golfer.

by Kyled85 on Mar 1, 2010 12:10 AM EST reply actions

Haven't Read Shit

But , simply put, if you average 5 RPG and your defense gives up an average of 4.5 RPG, the closer you stay to average the better off you are. The 1960 Yankees outscored the Pirates by 20+ runs in the Series, but they lost the Series 4-3. True, this involves pitching and defense, but what does it matter if you lose a lot of -10 run games if you average a consistent number of runs scored and it’s greater than runs allowed.

I used to be an A's fan until they left town and got good.

by philofthenorth on Mar 1, 2010 1:14 AM EST up reply actions

Extreme Example!

Royals score 700 runs in game 1. They get shut out for the next 161, and go 1-161. So, yes, I think this is fairly obvious. Scoring 1,8,1,8,1,8,1,8,1,8,etc. will probably net you around a .500 record. Scoring 4,5,4,5,etc, may get you more depending on your pitching. This is pretty simple.

by AxDxMx on Mar 1, 2010 1:48 AM EST up reply actions 1 recs

The Trouble Is

To my knowledge, no team has ever stayed even within a run of it’s average for a season. No research, it just seems impossible. I don’t think that stat even exists.

I used to be an A's fan until they left town and got good.

by philofthenorth on Mar 1, 2010 2:12 AM EST up reply actions

why is that trouble?

by kcdc1 on Mar 1, 2010 10:14 AM EST up reply actions

No Empirical Way

To test the theory.

I used to be an A's fan until they left town and got good.

by philofthenorth on Mar 1, 2010 4:10 PM EST up reply actions

I wasn't viewing this as a OBP vs Slug

I was viewing this as consistency of a player within a team. Like if a player always went 1 for 4 with a walk, vs. a player who had a good month and then two terrible ones.

I’m not sure how the OBP vs Slug got into it.

At least Wally Joyner's not on the team....

by tcon125 on Mar 1, 2010 10:45 AM EST reply actions

It is not about an individual player, it is a team dynamic.

Linear wieghts like Woba don’t take these into effect. You need a few high SLUG players to bring home your high OBP players, and Woba assumes that you will have a normal amout of high SLUG players on your team. Woba is still a lot better then OPS in determining value of the players, but it has some limitations in roster construction because it does not take into account how different types of hitters dynamically react to one another in scoring runs.
You can get 15 runners on base during a game and not score, but if you hit a homerun, you score. If you have people getting on base for the guy hitting homeruns, then you score more runs.

Go Royals!

by BabyBlues on Mar 1, 2010 12:13 PM EST up reply actions

How about

one good month and then (so far) two terrible years?

"Now…put that in your [BLEEP]ing pipe and smoke it." -Hal McRae

"I was doing this when BJ was in his father's nutsack." -Renzo Gracie

by Sweep_the_Leg on Mar 1, 2010 12:15 PM EST up reply actions

the thrust of his article

was that given the same average runs scored, higher SLG will cause a more favorable run distribution, resulting in more wins—even without more runs scored.

by kcdc1 on Mar 1, 2010 12:15 PM EST up reply actions

Tango has a thread on it

This AM

So it looks like the “consistency” part isn’t all that novel. Also sounds like the OBP/SLG claim is problematic, but I’ll leave this to the real math people.

I'm not a sabermetrician, but I do play one at FanGraphs.

Can't get enough of me? Check out my Twitter feed.

by Matt Klaassen on Mar 1, 2010 11:14 AM EST reply actions