OBP....and its importance

[TyCobb]

Um, interleague play, late in the game, empty bench, P on deck, tie game or down by 1, guy on 3B, 1 out... take the Sac Fly or run-scoring GB out...

I would take the walk.

[El Gordo]

I would take the walk.

What would TY Cobb do?:laughlol:

[trk001]

Jump into the stands and beat a cripple.

[now]

A good read about the correlation of different statistics with Runs Scored with a shiny graph, followed by an article on sabermetric stats for hitters in general.

http://www.hardballtimes.com/main/article/ops-for-the-masses/

I agree, OBP is most important.

That's a fascinating chart, and it's interesting how all the fancy-schmancy offensive stats really add very little to what you get out of OPS. But this seems inconsistent with the well-accepted notion that OBP is more important than SLG. Or is this chart not really addressing that point.

It seems that the new dogma around here (for the last couple of years) is "OBP is more important than SLG" and even "OBP is more important than OPS." The chart clearly shows that this assumption is false. Consider...

* SLG correlates with runs slightly better than OBP does.

* OPS correlates with runs much better than either OBP or SLG.

* OPS' works only slightly better than OPS while weighting OBP and SLG correctly.

* OPS is essentially as good as any other more advanced metric to correlate with runs.

[trk001]

* SLG correlates with runs slightly better than OBP does.

A .003 difference is statistically insignificant and is basically "white noise" in the data.

Also, OBP and SLG measure different things, one's not inherently more important than the other, and, as noted earlier in the thread, the two go hand in hand. Guys with good discipline get better pitches to hit and higher SLG ... guys with good power are pitched around and get higher OBP.

[Cavspider]

Walks are important not only for merely OBP but because they always mean a higher pitch count for the opposing pitcher.

It seems that a primary goal of winning teams is patience, not only to get a walk or better pitch to hit but to wear out the opposing pitcher with a high pitch count asap and get to the bullpen. Going deep into counts, even those at bats resulting in outs is important for this purpose.

It would be interesting to see the utility of someone who rarely walked but whose at bats lasted forever.

Say a guy rarely walked and hit .300 but every at bat that produced an out was at least 10 pitches. If every guy on the team could do that youd wear out the opposing starter really fast.

[mweb]

Walks and OBP are obviously very important, and is why we should have gotten Johnson over Atkins.

[now]

A .003 difference is statistically insignificant and is basically "white noise" in the data.

Also, OBP and SLG measure different things, one's not inherently more important than the other, and, as noted earlier in the thread, the two go hand in hand. Guys with good discipline get better pitches to hit and higher SLG ... guys with good power are pitched around and get higher OBP.

Sure, my point was not to tout SLG per se but to debunk the "OBP is better" theory. The two measures and talents do go hand in hand.

While it might be true that in the statistical context of the advanced calculation of OPS' "OBP is more important," when it comes to actual impact on the game - runs created - we can see from the chart that actually OBP is less important (slightly) than SLG and clearly less important than OPS.

[Frobby]

It seems that the new dogma around here (for the last couple of years) is "OBP is more important than SLG" and even "OBP is more important than OPS." The chart clearly shows that this assumption is false. Consider...

* SLG correlates with runs slightly better than OBP does.

* OPS correlates with runs much better than either OBP or SLG.

* OPS' works only slightly better than OPS while weighting OBP and SLG correctly.

* OPS is essentially as good as any other more advanced metric to correlate with runs.

Clearly OBP is not better than OPS. That's why people started using OPS in the first place, after it was found to be better than either OBP or SLG by itself.

However, I'm not quite sure whether the fact that SLG seems to correlate with runs slightly better than OBP means that SLG is more important than OBP. I'm pretty sure there have been a number of simulations that show OBP is more important than SLG, by factors ranging between 1.7 and 3 depending who you listen to.

[trk001]

Neither OBP nor SLG can be "less" important than OPS because OPS is OBP+SLG. It's not that OPS is more or less important, it's that it correlates better to runs scored, that's all.

[Old#5fan]

No, it is as good as a hit. Because the walk puts more men on base and makes it more likely taht the person behind that person will score a run.

Notice when the bases are empty a BB is statistically equal to hit. The only thing that allows a single to be a greater run-producer than a BB is the fact that a league average player should be able to score from second on most singles.

Notice the point of the graph, the more men on base - the more likely you are to score runs any number of ways.

Therefore being able to take a walk is a very important part of the game.

How pray tell does a "walK' put more men on base than a "hit" and make it more likely the person will score a run? That is simply nonsensical and proves beyond any measure you have no idea what you are talking about. A hit also puts a runner on and often times not just on first base Mr. Baseball Einstein!:laughlol:

[amateurfan]

That's a fascinating chart, and it's interesting how all the fancy-schmancy offensive stats really add very little to what you get out of OPS. Bu this seems inconsistent with the well-accepted notion that OBP is more important than SLG. Or is this chart not really addressing that point.

I thought it was pretty well established the other way round.

Edit: Not that I'd know and I was thinking of the correlation only, not the cause.

[wildcard]

Lots of talk around here about needing a big bat...I think we should be adding OBP moreso.

Let's look at some rankings from the last few years..OBP and runs scored(AL only):

2009:

1) Yankees...1st in OBP...1st in runs scored.

2) Boston...2nd in OBP...3rd in runs scored.

3) Angels....3rd in OBP...2nd in runs scored.

4) Twins...4th in OBP...4th in runs scored.

5) Tampa...5th in OBP...5th in runs scored.

6) Baltimore...8th in OBP...11th in runs scored.

2008:

1) Boston...1st in OBP....2nd in runs scored.

2) Texas..2nd in OBP...1st in runs scored.

3) Yankees...3rd in OBP...7th in runs scored.

4) Detroit...4th in OBP...4th in runs scored.

5) Minnesota...5th in OBP...3rd in runs scored.

6) Baltimore...8th in OBP...8th in runs scored.

2007:

1) Yankees...1st in OBP...1st in runs scored.

2) Boston...2nd in OBP...3rd in runs scored.

3) Angels...3rd in OBP...4th in runs scored.

4) Detroit...4th in OBP...2nd in runs scored.

5) Cleveland...5th in OBP...6th in runs scored.

6) Baltimore...9th in OBP...9th in runs scored.

Pretty telling stuff there...Moral of the story? Get on base and you score runs. Its that simple.

I think the focus is much too narrow here.

If you look at the last 5 years the only two stats that are common with most runs scored are 1) Hits and 2) OPS

2009- Yankees (915 runs) led in H, HR, RBI, OBP, SLG, OPS

2008- Texas (901 runs) led in H, 2B, TB, RBI, BA, SLG, OPS

2007- Yankees (968 runs) led in H, HR, TB, RBI, BA, OBP, SLG, OPS

2006- Yankees (930 runs) led in H, RBI, BA, OBP, OPS

2005- Boston (910 runs) led in H, 2b, RBI, BA, OBP, OPS

1) First thing to realize is teams that lead the league in runs lead in a lot of categories.

2) In 2008 Texas scored the more runs then Boston (910 to 845) but was 4 points lower in OBP (354 to 358). That is a 56 run difference.

3) The team with the most hits and highest OPS, has scored the most runs every year for the last 5 years.

http://espn.go.com/mlb/stats/team/_/stat/batting/year/2008/league/al

[KAZ97]

Clearly OBP is not better than OPS. That's why people started using OPS in the first place, after it was found to be better than either OBP or SLG by itself.

However, I'm not quite sure whether the fact that SLG seems to correlate with runs slightly better than OBP means that SLG is more important than OBP. I'm pretty sure there have been a number of simulations that show OBP is more important than SLG, by factors ranging between 1.7 and 3 depending who you listen to.

I ran into this issue when playing around with the data. To properly look at correlations like the ones present above, you need to start with individual, plate appearance level data. To look simply at aggregate statistics invites problems, such as the ecological fallacy.

The problem with dealing with the individual data is that not all events are defined for both OBP and SLG. For example, what is the OPS of a batter that has 3 plate appearances, two walks and a sac fly? The correct answer is undefined, since formally SLG is undefined for those events. If one assumes SLG (or BA) is 0.000, then you will get wrong results when aggregating.

OPS is a good, back of the napkin estimation of talent because in aggregate it correlates with runs scored. But using it, or comparing it's components, in any stringent manner is difficult.

[now]

Clearly OBP is not better than OPS. That's why people started using OPS in the first place, after it was found to be better than either OBP or SLG by itself.

However, I'm not quite sure whether the fact that SLG seems to correlate with runs slightly better than OBP means that SLG is more important than OBP. I'm pretty sure there have been a number of simulations that show OBP is more important than SLG, by factors ranging between 1.7 and 3 depending who you listen to.

Maybe this is just semantics ... but let's try to clarify. In essence OBP and SLG are virtually equal in "importance" in the context of correlation to runs created (according to the chart). The factors you mention, of different weights (1.8 cited in the article the chart came from), really are technical adjustments because of different denominators, double counting, etc. To me the "more important" issue is which metrics correlate best to runs created.