Value of OBP, Relative to SLG

[rluzinski]

In another thread, my contention was that 1 point of OBP was worth more than 1 point of SLG. Since that notion was roundly rejected I felt in neccessary to prove my point. Using Multiple Linear Regression, I was able to find the relative weights of OBP and SLG for prediction runs scored:

 SLG OBP RUNS LIN REG Boston Red Sox 0.451 0.357 479 475 New York Yankees 0.449 0.358 486 474 Baltimore Orioles 0.468 0.334 436 457 Los Angeles Angels 0.418 0.326 423 400 Florida Marlins 0.415 0.338 390 415 Chicago Cubs 0.451 0.327 410 431 San Francisco Giants 0.407 0.331 397 398 St. Louis Cardinals 0.427 0.341 447 430 Texas Rangers 0.476 0.330 476 458 Detroit Tigers 0.420 0.323 396 398 Tampa Bay Devil Rays 0.409 0.326 402 392 Minnesota Twins 0.409 0.334 398 404 Colorado Rockies 0.415 0.332 389 406 Toronto Blue Jays 0.413 0.333 428 406 Philadelphia Phillies 0.411 0.348 423 426 Oakland Athletics 0.396 0.336 406 395 Cincinnati Reds 0.442 0.337 434 438 Kansas City Royals 0.398 0.321 387 375 Atlanta Braves 0.432 0.330 431 419 Los Angeles Dodgers 0.407 0.332 387 399 San Diego Padres 0.404 0.338 406 405 Chicago White Sox 0.418 0.322 414 395 Cleveland Indians 0.431 0.323 406 408 Washington Nationals 0.395 0.328 359 383 Arizona Diamondbacks 0.419 0.331 400 409 Seattle Mariners 0.389 0.317 380 362 New York Mets 0.410 0.321 393 386 Milwaukee Brewers 0.418 0.331 396 408 Houston Astros 0.405 0.319 365 379 Pittsburgh Pirates 0.402 0.321 367 379 

 

http://home.new.rr.com/mffl03/risp.jpg

 

The equation I found was:

 

Runs = 1444 x OBP + 898 x SLG - 443

 

1444 / 898 = 1.6

 

One point of OBP is equal to 1.6 points of SLG.

 

EDIT: Opps, thanks endhttp://forum.brewerfan.net/images/smilies/smile.gif Also, what I did here has been done before, incase someone thinks I'm trying to say I "invented" this.

[endaround]

Well except your last line is wrong...

[Monty]

Is this supposed to be a comparison between OPS and SLG or OBP and SLG? Maybe I am confused.

[rluzinski]

DOH! I have a REAL problem mixing up OBP and OPS!!

 

It's comparing OBP to SLG http://forum.brewerfan.net/images/smilies/frown.gif

[1992casey]

rluzinski,

 

I understand your point, but I'm lost with "Multiple Linear Regression"! http://forum.brewerfan.net/images/smilies/happy.gif

 

Using simple logic (which, of course, doesn't always work) it would stand to reason that a point of OBP is worth more than a point of SLG. OBP covers 1,000 points while SLG covers 4,000.

 

Still, it's interesting that OPS works so well, better than OBP or SLG individually.

[rluzinski]

OBP covers 1,000 points while SLG covers 4,000.

 

I'm sure that's part of it, but with the fascination with OPS, it's good to realize all OPS are not created equal. I fell victim to that when I didn't give Clark his props for having a high OBP.

 

Hardball times uses a stat called GPA:

 

"Gross Production Average, a variation of OPS, but more accurate and easier to interpret. The exact formula is (OBP*1.8+SLG)/4, adjusted for ballpark factor. The scale of GPA is similar to BA: .200 is lousy, .265 is around average and .300 is a star."

 

They say 1.8, so far this year it's been 1.6, but the idea remains the same; comparing just OPS's of two players isn't really fair since OBP is that much more important.

[shaqkobefan]

This is one of the most rediculously umpired games I have seen in a while. Livan Hernandez is not that good of pitcher. Every pitch off the outside corner is a strike. It's rediculous. If we lose this game by one run it should be protested.

[shaqkobefan]

Sorry, wrong thread... my bad guys.

[rluzinski]

Wrong thread but I agree, I have NO idea what a strike is anymore.

[Picciolo]

Quote:

---

The equation I found was:

 

Runs = 1444 x OBP + 898 x SLG - 443

 

1444 / 898 = 1.6

 

One point of OBP is equal to 1.6 points of SLG.

---

You can't take the ratio of two coefficients of different scales like this. What were the p-values on each?

[rluzinski]

You can't take the ratio of two coefficients of different scales like this. What were the p-values on each?

 

Umm, yeh, I'm an engineer, not a statistician http://forum.brewerfan.net/images/smilies/smile.gif The r^2 value is .79, whish isn't terribly high, but good enough to estimate like I was trying to do. I used the LINEST function in excel, if that helps.

 

I don't understand why you say the scales need to be the same to compare their effect on runs. I'm certainly not arguing that I am right, just that when I got an answer that made sense, I ran with it. I've read many times that OBP usually coorelates to runs with a coefficients between 1.5 and 1.8.

 

EDIT: I didn't expect the r^2 value to be too high because I know treating all hits as qual to 1 is wrong. Runs Created or Linear Weights estimates runs scored better.

[Picciolo]

R-squared's run from 0 to 1. A 1 would indicate that runs could be explained entirely by OBP and SLG. .79 is actually very high considering you're only looking at two variables. When you are running multivariate regression models, each variable will have a "p-value". Some software packages may report only a t-score. A very low p-value (<.05) would indicate that a variable is significantly related to runs. A high t-score (>1.96) indicates the same thing. We can start to compare relative importance to scoring runs by looking at the relative p-value (t-value) of OBP and SLG. There are other issues to consider, but this is at least a start.

 

The reason you can't compare two coefficients of different scales is for the following reason:

 

Example (forget the intercept term): Runs=2*OBP + 1.5*SLG

 

If you're looking at a team that has OBP of 300 and SLG of 400 you get 2*300+1.5*400=1200. In other words, they contribute equally to runs.

 

When I use Excel for regression, I use tools>internet options>regression.

[rluzinski]

-squared's run from 0 to 1. A 1 would indicate that runs could be explained entirely by OBP and SLG. .79 is actually very high considering you're only looking at two variables.

 

That much I knew. Why are you suprised that OBP and SLG would predict runs so well? What other variable should I have in their, "battleness"?

 

If you're looking at a team that has OBP of 300 and SLG of 400 you get 2*300+1.5*400=1200. In other words, they contribute equally to runs.

 

I'm not sure exactly what you mean, but perhaps I didn't make it clear what I was trying to do. My point was all OPS of .800 are not created equal:

 

.300 OBP .500 SLG:

 

Runs = 1444 x .300 + 893 x .500 - 444 = 436 runs

 

.400 OBP .400 SLG:

 

Runs = 1444 x .400 + 893 x .400 - 444 = 491 runs

[Picciolo]

Quote:

---

Why are you suprised that OBP and SLG would predict runs so well? What other variable should I have in their, "battleness"?

---

LOL! Actually, I was thinking that there is a lot more that goes into run-scoring than OBP and SLG, such as speed, weather, ballpark, opposing pitcher, league, manager style, and yes, battleness. Actually, I think the PC term is grindingness, but now we're splitting hairs.

 

Quote:

---

I'm not sure exactly what you mean, but perhaps I didn't make it clear what I was trying to do. My point was all OPS of .800 are not created equal:

---

OK. I did miss that one. I'll have to think about that one some more when I don't have a few pale ales in my system. http://forum.brewerfan.net/images/smilies/smile.gif

[Loopee]

Quote:

---

I'm not sure exactly what you mean, but perhaps I didn't make it clear what I was trying to do. My point was all OPS of .800 are not created equal:

---

 

Mission complete.

 

Well, at least you've convinced me.

 

Because of the different scales though, I don't know how much this means outside the context of OPS. Any way to predict whether a guy who has relatively high (to the other players in the league) OBP or SLG will help the team more?

 

Like would you rather have the 3rd best OBP or the 3rd best SLG?

[1992casey]

---

I have NO idea what a strike is anymore.

---

I don't know what traveling means in the NBA anymore, either.

[Loopee]

Quote:

---

I don't know what traveling means in the NBA anymore, either.

---

 

At least you know that refs call traveling and umps call strikes.

 

That puts you ahead of most people on the mlb.com forums.

[rluzinski]

Like would you rather have the 3rd best OBP or the 3rd best SLG?

 

What I'm saying is 10 OBP points = ~16 SLG points. Clark, as a result, is more valuable this year than his OPS indicates.

[endaround]

The biggest problem with running this regression isn't using the ratio of paramters--that just a ratio of the means. Now each parameter will have a different variance attached to it meaning the ratio's distribution isn't cut and dry, but thats secondary here to some degree (though proving the ratio's CI doesn't include numbers less than 1 is really what Russ should be aiming for). The biggest problem is that OBP and SLG are highly correlated (its called BA) which will bias the parameters.

[strawbossisevil]

I'd just like to say that I dislike math.

 

That will be all.http://forum.brewerfan.net/images/smilies/pimp.gif

[buddy79]

you stat nerds need to get out of baseball. back to the chess board back!! your ruining our game!

[rluzinski]

Well, I appreciate that they aren't trully 100% independant variables and that a better function of run scored would use the product of the two numbers, not the sum. I will be the first to admit that I am not a statistician, but my aims weren't to prove that the formula I dervied was the "right" one. While you can debate about the exact coefficients that should be used, it's clear that OBP will drive the runs function more than SLG.

 

Again, this was JUST to show the OBP of OPS was more important. That's all, I swear!

[Bruce Ciskie]

OK...all those numbers made my head hurt.

 

A friend of mine coaches college baseball up here. He says you should double OBP and add it to SLG. He needed about ten minutes to sell me on the idea, so you guys aren't that far off with your mathematical stuff here.

 

I'm going to chat about it with Rob Neyer at some point and get his thoughts. It's an interesting hypothesis.

[rluzinski]

I'm sure Neyer has seen others do this already, it's not new at all. True SABR guys have alredy sworn off OPS and OxS anyway, since it weights types of hits (single, double, etcs) as 1, 2, 3, etc... this just isn't right if your aim is to gauge a player or team's ability to create runs. I like the linear weights method, personally. You can read more on it at tangotiger.net.

[danzig6767]

If OBP is nearly TWICE as important than slugging how is

 

Texas third in runs scored and 18th in OBP and 1st in SLG?

 

White Sox/Texas finish 3rd/4th in runs last year but finish 15th/21st in OBP?

 

This year the top 8 teams in SLG% are all in the top 8 in runs except Atlanta who is 9th in runs.

 

Three of the top 8 OBP teams are Florida/San Diego/Oakland,in runs they rank 12th/13th/21st in runs.Florida is 5th in OBP and 21st in runs.

 

In the last three years if you looked at the top 10 teams in runs scored,24 of those 30 teams were also in the top 10 in SLG%.Two of those that missed were 11th in runs.

 

In the same 3 year span 23 teams in the top 10 in runs were also in the top 10 in OBP.

 

In the last three years,the 10 worst SLG teams finished among the 10 worst scoring teams 27 out of 30 times regardless of their OBP numbers.

 

Good luck trying have a very prolific offense without a high SLG%,even if the team gets on base well.If OBP is so much more important than SLG,then why dont the rankings in runs scored show a higher correlation?Sure seems to me where a team ranks in SLG compared to other teams will play as big or nearly as big a role as OBP in how high they rank in runs scored.