SLG% and stolen bases

[rousieboy]

Anyone,

Quick question...does SLG% take into account stolen bases? If not, isn't a stolen base from 1st to 2nd like hitting a double? 2nd to 3rd like hitting a triple?

 

I like OBP, SLG, and OPS...but was just wondering about this point...thanks Rousie

[sbrylski]

Quick question...does SLG% take into account stolen bases? - No.

 

If not, isn't a stolen base from 1st to 2nd like hitting a double? 2nd to 3rd like hitting a triple?

 

Also no. You can't advance other runners with a steal. And, you can get caught stealing.

 

A single + stolen base is roughly only worth 80% of a double. A double and stolen base is roughly worth 85% of a triple.

 

So you can't just add it into their SLG, you'll lose accuracy. And stolen bases are much less valuable than most would think, unless you very, very rarely get caught. So that's why OPS correlates so well with runs scored despite ignoring SB and CS.

[rluzinski]

If you remove a single for every CS and add a base for every SB, I don't think it would be the worst way to try and combine the two. You are still going to over value the average SB, though.

 

Remember, though, a stolen base attempt is different than most batting events since a runner can decide when to try one. Context is absolutely critical (base/out/inning/score/batter/pitcher) in determining the potential value of a SB and the potential cost of a caught stealing. Here, you can see how the base/out situation affects the run value of a SB and CS:

 

http://www.tangotiger.net/RE9902event.html

[And That]

I would not have guessed that the worst time to be caught stealing or picked off was going from 1st to 2nd with one out and a runner on third. Makes sense I guess.

 

That's a really interesting chart. Thanks for sharing that.

 

Edit: After looking at it again, did I understand the situation correctly? Does the chart mean the "CS" is the guy going from 1st to 2nd, the guy going from 3rd to home, or both?

[TooLiveBrew]

Context is absolutely critical (base/out/inning/score/batter/pitcher) in determining the potential value of a SB and the potential cost of a caught stealing.

 

Hall's SB in the 9th last night is probably a good example of this. The result of him being safe was very big, but imagine if the P had stepped off the rubber there -- the negative impact would've probably been even bigger. Not trying to suggest Hall just got lucky -- clearly he read the situation very well, but it was still a huge gamble when you consider that he was already in scoring position.

 

 

EDIT: From checking the chart rluz shared, the safe SB was a .38, but a CS there would have been a -.60. Nearly double the value, but in the negative direction! Yikes. Boy am I glad Hall was safe.

[Brian the Automator]

If you remove a single for every CS and add a base for every SB, I don't think it would be the worst way to try and combine the two. You are still going to over value the average SB, though.

 

Remember, though, a stolen base attempt is different than most batting events since a runner can decide when to try one. Context is absolutely critical (base/out/inning/score/batter/pitcher) in determining the potential value of a SB and the potential cost of a caught stealing. Here, you can see how the base/out situation affects the run value of a SB and CS:

 

http://www.tangotiger.net/RE9902event.html

My problem with valuing stolen bases and speed in general is the pressure it puts on the pitcher. Is the pitcher going to be less apt to throw a breaking ball when the delivery to the plate needs to be quicker?... Does it take pitchouts and the free ball it gives the batter into account? I'm really not sure how you can formulate that... I don't think it's a HUGE part of the game, but I don't think it gets enough credit in statistical circles.

[TooLiveBrew]

Clearly those kinds of intangibles are already measured in W-L record & clutchiness. Your stats don't impress me, Brian. http://forum.brewerfan.net/images/smilies/wink.gif

[rluzinski]

My problem with valuing stolen bases and speed in general is the pressure it puts on the pitcher. Is the pitcher going to be less apt to throw a breaking ball when the delivery to the plate needs to be quicker?... Does it take pitchouts and the free ball it gives the batter into account? I'm really not sure how you can formulate that... I don't think it's a HUGE part of the game, but I don't think it gets enough credit in statistical circles.

 

Studies that have tried to make all the necessary adjustments and have come to a surprising situation. On average, whatever the negative effect that basestealer has on the pitcher's performance is apparently not as large as the negative effect that it has on the batter. If I remember correctly, having any base runner at 1B helps the batter by opening a hole (especially for left handers), so that adjustment is made. when you just focus on the base stealing specialists, the average batting performance goes down.

 

Now, I can't seem to find that article but here's one that concludes the effect is probably real but very small:

 

http://www.hardballtimes.com/main/article/base-stealer-intangibles-part-1/

 

http://www.hardballtimes.com/main/printarticle/base-stealer-intangibles-part-2/

 

Of course, that's "on average". I'm sure there are some pitchers that suffer more on average (although it might be hard to determine who because of sample size issues) and I'm sure there are some batters who could give a darn (did i read that vets ignore base runners on average)? Either way, we are talking about a run here or a run there. Not as large as common sense might suggest.

 

I hope someone else is aware of the studies I recall but can't find.

[rluzinski]

Edit: After looking at it again, did I understand the situation correctly? Does the chart mean the "CS" is the guy going from 1st to 2nd, the guy going from 3rd to home, or both?

 

I would think it's any "CS" event occurring in that situation, so both. I suspect that most are the 1st to 2nd variety. You should be able to check those numbers using this:

 

http://www.tangotiger.net/RE9902.html

 

1.243 (1 & 2,1 out) - .387 (3, 2 out) = .856 (compared to the .95 of the previous table). Probably a few of those CS were the 3-H variety, which would be more costly?

EDIT: From checking the chart rluz shared, the safe SB was a .38, but a CS there would have been a -.60. Nearly double the value, but in the negative direction! Yikes. Boy am I glad Hall was safe.

 

Really, in that case, RE is pretty inconsequential. You only need 1 run. You are better off looking this:

 

http://www.tangotiger.net/RE9902score.html

 

Runner on 2B with 1 out scores at least 1 run 41% of the time.

Runner on 3B with 1 out scores at least 1 run 66% of the time.

No runner and 2 out scores at least 1 run 8% of the time.

 

Turn that into win%:

 

Chance of winning before SB: 71% (50% + .5 x 42%)

Chance of winning after SB: 83% (50% + .5 x 66%)

Chance of winning after CS: 54% (50% + .5 x 8%)

 

That agrees with the actual results pretty well:

 

http://winexp.walkoffbalk.com/expectancy/search/?start_year=1977&end_year=2006&team=H&inning=9&outs=1&expectancy[bases]=2&scorediff=0

 

A SB adds 12% to the win%, a CS subtracts 29%. If I'm doing to right, a CS is 2.4 times more costly than a SB is beneficial. You better be darn certain your SB attempt is going to succeed in that situation.

[TooLiveBrew]

A SB adds 12% to the win%, a CS subtracts 29%. If I'm doing to right, a CS is 2.4 times more costly than a SB is beneficial. You better be darn certain your SB attempt is going to succeed in that situation

 

Yeah, it was a huge gamble by Hall. He's both 'good' & 'incredibly lucky' in that sense.

 

I has to ax you a kwesshin, rluz --

 

Turn that into win%:

 

Chance of winning before SB: 71% (50% + .5 x 42%)

Chance of winning after SB: 83% (50% + .5 x 66%)

Chance of winning after CS: 54% (50% + .5 x 8%)

 

What are the 50% & .5 in those equations? I'm asking this knowing full well that the answer is probably mind-numbingly simple. I looked at the tangotiger link you provided, but couldn't figure out those two numbers. Thanks in advance.

[olliegjw]

Quick question...does SLG% take into account stolen bases? - No.

 

If not, isn't a stolen base from 1st to 2nd like hitting a double? 2nd to 3rd like hitting a triple?

 

Also no. You can't advance other runners with a steal. And, you can get caught stealing.

 

A single + stolen base is roughly only worth 80% of a double. A double and stolen base is roughly worth 85% of a triple.

 

So you can't just add it into their SLG, you'll lose accuracy. And stolen bases are much less valuable than most would think, unless you very, very rarely get caught. So that's why OPS correlates so well with runs scored despite ignoring SB and CS.

 

Wouldn't the same arguments be true for walks? Other runners don't advance, and they are not as valuable as a single. This is why I feel some people overvalue players like Gabe Gross. His OBP is okay because of his BB numbers, but his average still stinks.

[sbrylski]

Wouldn't the same arguments be true for walks? Other runners don't advance, and they are not as valuable as a single. This is why I feel some people overvalue players like Gabe Gross. His OBP is okay because of his BB numbers, but his average still stinks.

 

A walk is only worth ~2/3rds of a single, but its better to count them and look at OBP than use BA and not count them at all. And of course you'd rather have an OBP made up 100% of batting average, but it would take a fairly large difference in batting averages to make up for even a small difference in OBP.

 

And OBP really is just what it says - how often a player gets on base. It does not attempt to assign a weight to how the reach base. So that's why you add in the SLG, which does assign a weight to the quality of the players hits, and despite still ignoring the difference between a single and a walk in the OBP component, it correlates 96% to scoring runs.

 

And if you think about that, a hit is counted both in OBP and SLG, while a walk just helps OBP. So in OPS, a walk may actually be undervalued.

[TooLiveBrew]

Why don't you go marry these 'stats', huh?! http://forum.brewerfan.net/images/smilies/wink.gif

[rluzinski]

Wouldn't the same arguments be true for walks? Other runners don't advance, and they are not as valuable as a single. This is why I feel some people overvalue players like Gabe Gross. His OBP is okay because of his BB numbers, but his average still stinks.

 

To echo sbrylski, no one is suggesting that OBP captures all the value of a player. Fortunately, SLG captures the other side of the coin, as it's measuring the total bases a player gets via a hit (not walk). That's why, for all it's crudeness, OPS (OBP + SLG) does a reasonable job of estimating a player's offensive production.

 

For a better explanation for why OPS works, read this excellant article:

 

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

 

The author, Dan Fox, now works for the Pirates.

[TooLiveBrew]

What is it about EQA that makes it (iirc) more accurate than OPS?