More valuable base stealer: Kinsler of Ellsbury?

Just for steals,

I’d rather have Ellsbury

Remember: baseball guys... baseball...

by Metty5 on Oct 7, 2008 12:55 PM EDT reply actions 0 recs

I'd multiply CS by 3 then subtract from SB

to get “net steals”.

Kinsler actually comes out slightly ahead on that metric. However, that assumes the break-even point is 75% when it’s actually more like 72%. So I’d call it very, very close to even.

Your 2008 Athletics: It's Nothing Personal.

by PaulThomas on Oct 7, 2008 1:39 PM EDT reply actions 0 recs

This assumes that the SB attempts are taken at arbitrary situations, but that’s not the case. The offense has the choice of when to attempt the steals. Since some situations change the break-even point to as low as 50% for break-even (or less for double-steals in certain situations), and one should assume that – using normal “game theory” strategies – the offense will choose to steal in situations where the percentages are balanced in their favor.

by BobbyMac on Oct 8, 2008 3:47 AM EDT up reply actions 0 recs

I'm not sure why you'd assume that, given that baseball is played by baseball players

and managed by baseball managers.

Also, while a lower “true break even point” would weigh more in Ellsbury’s favor, the returns are diminishing. There are only so many high-leverage situations you can steal in. The more often you steal overall, the more your break-even point will converge with the average.

Your 2008 Athletics: It's Nothing Personal.

by PaulThomas on Oct 8, 2008 3:46 PM EDT up reply actions 0 recs

Ellsbury....

Stealing twice as much means he is attempting them more often, which not only helps the team because you are getting to 2B, but you are also distracting the pitcher. If Ellsbury was getting caught more than 20% of the time, I think Kinsler would have a much better case here, But stealing 100% more bases and only getting caught 8% more. I’d go with Ellsbury pretty handily.

"My mom always taught me it's better to laugh at yourself than to laugh at others. She was so wrong. ;)" -Pedrophile

by Boxkutter on Oct 7, 2008 4:46 PM EDT reply actions 0 recs

Having a base-stealer on first HELPS the pitcher

when you actually do the research.

Hitters are distracted more than pitchers.

Your 2008 Athletics: It's Nothing Personal.

by PaulThomas on Oct 7, 2008 7:03 PM EDT up reply actions 0 recs

I would assume

I’d assume that this depends on the runner and the hitter.

by Fanon on Oct 8, 2008 12:49 AM EDT up reply actions 0 recs

i've always believed this....

….but i’m curious how they went about researching it.

did they compare hitting stats with “basestealers” versus “non-basestealers” on first? (and, if so, how did they define the two categories?)

you can’t really compare pitching stats with and without a runner on first, because i assume that all pitchers pitch worse when there’s a baserunner on, not because they necessarily are distracted (or because they’re worse from the stretch — though this is a big factor too), but because any pitcher who has allowed a baserunner to reach will, at the margins, be less likely to have his best stuff at the moment. (for the extreme example, consider situations where the pitcher has loaded the bases — clearly, he doesn’t have his best stuff at the moment.)

anyway, just curious what the method was.

by bleedjaxblue on Oct 8, 2008 2:11 AM EDT up reply actions 0 recs

Just read "The Book"...

every devoted baseball fan should basically read it anyway at some point. I’m sure you can get it on interlibrary loan if you don’t feel like lining Tom Tango’s pockets.

Your 2008 Athletics: It's Nothing Personal.

by PaulThomas on Oct 8, 2008 2:17 AM EDT up reply actions 0 recs

hmmm

no chance of getting the distilled version for those of us who have neither the time nor the interest in reading an entire book to get the answer to one specific question?

i’m interested in what Tango’s method is here. not in Tango in general.

by bleedjaxblue on Oct 8, 2008 2:52 PM EDT up reply actions 0 recs

The Book...

… is mostly a presentation of runs-by-situation charts, and win-by-situation charts. B-R has added win-by-situation to their game logs, based on the data from The Book, and walkoffbalk.com has these also. BP has annual runs-by-situation data.

From these averages, they draw some conclusions, and show how much various tactics impact run scoring (and thus expected win%). For example, there’s a good presentation on the tactic of batting the pitcher 8th.

Clearly, this is a very broad brush synopsis, and I do agree with PaulThomas that it should be in any fan’s library. Even if many of the conclusions are overly simplistic, the numbers are good to have around, and the simplistic conclusions are a good starting point for any discussion such as this one. For example, if one were to assume that SB attempts were made in a vacuum, a team (or player) would need to succeed 72% of the time (as noted above) to break even. So, if a team (or player) isn’t succeeding at that rate, there either need to be good situational explanations for the attempts, or the team (player) needs to consider being less aggressive.

by BobbyMac on Oct 8, 2008 3:37 PM EDT up reply actions 0 recs

cool....thanks

so, i guess what my question is, what was the baseline for the above situation?

because it makes a big difference (at least i’d guess) whether you’re comparing “hitters stats for man on first v. no man on first” or “pitchers stats for base-stealer on first vs. nonthreat on first” or some other permutation of that (or something else entirely).

as i said, my intuition agrees with this result — just want to know how he got there, since i’ve been curious about a way to “prove” this for a while.

by bleedjaxblue on Oct 8, 2008 3:55 PM EDT up reply actions 0 recs

Yeah, I just don't remember

I’ll try to look it up in the next couple of days.

Your 2008 Athletics: It's Nothing Personal.

by PaulThomas on Oct 8, 2008 7:19 PM EDT up reply actions 0 recs

appreciated

by bleedjaxblue on Oct 9, 2008 12:05 AM EDT up reply actions 0 recs

Time played...

Are these stats counting the fact that Kinsler has missed a majority o the season?

by team name deleted on Oct 7, 2008 5:21 PM EDT reply actions 0 recs

He didn't miss a majority of the season

Missing 41/162 is not a majority

You would actually need more than half, that would mean 82, to miss the majority of the season.

Either way, it’s unlikely he improves on his success rate so the comparison is probably still valid.

by Kenan and Kel on Oct 7, 2008 6:01 PM EDT reply actions 0 recs

Thanks for the opinions, I think they are all good. But here’s the thing, SB and CS are not all created equally, right? For example, if you are the back end of a double steal, that shouldn’t be weighted as heavily as, say, stealing a base when your team is tied in the 9th. Likewise for CS, if you are caught because the hitter missed on a hit and run, that shouldn’t be valued equally with a normal CS.

I don’t have the statistical prowess that some others due on this blog, but I’d be curious if anyone had any ideas of how to come up with a way of valuing these different kinds of SB/CS. Or am I wrong, and are all SB and CS the same?

by knuckleballer on Oct 8, 2008 11:24 AM EDT reply actions 0 recs

I think (for the most part) the point of a stolen base is to get a runner into scoring position, in total isolation stealing second base will obviously do that for you. At the same time, hitting a double, triple or homer obviously does that just as well, if not better. Ellsbury played about 16% more games than Kinsler this year, if you project Kinsler’s numbers out over 145 ABs (equal to Ellsbury) you see he should have managed to get himself into scoring position (or all the way home) with XB hits and steals roughly 107 times. Ellsbury only pulled that off 88 times. There has to be something to be said for the number of times Kinsler didn’t need to steal second because he was already standing there thanks primarily to his bat.

That being said – which player would be more valueable without any steals at all? I think there is no doubt that the player you would want in that case is Kinsler – Ellsbury might not even be able to hold down a major league job without the speed, let alone start, so to me I think the question you ask is pretty simple – Ellsbury’s speed makes him a productive starter for the Sox, Kinsler’s speed only adds to a good overall package for the Rangers. You gotta pick Ellsbury’s steals as more important to the Sox.

by Curveball 2000 on Oct 8, 2008 11:41 AM EDT reply actions 0 recs

Not the question

Of course kinsler is the better offensive PLAYER. That is fairly obvious though your analysis is valid. It was JUST a question of who’s STOLEN BASE record was more valuable. Not a lot to do with the particular players involved really.

by casejud on Oct 9, 2008 12:30 PM EDT up reply actions 0 recs

Actually, I think I answered the OPs first question – Ellsbury’s stolen bases are more valuable to the Sox because without the stolen bases he is at best a bench player and with them he is a good starter. Kinsler is a rock solid starter for the Rangers with or without the stolen bases. The first question the OP asked very specifically asks about Kinsler and Ellsbury. (Which player’s stolen bases are more valuable to his respective team, Ian Kinsler’s 26 steals at a 92% success rate or Ellsbury’s 50 steals at 82%?)

The question you say I didn’t answer (and you are right) is the second question the OP asked. (And what is more valuable for a team, a player stealing 25-30 bases at 90% or stealing 50-55 at 80%?)

I think my answer to that question would be a variation on the same theme – if the player stealing 25 bases and the guy stealing 50 both had the same number of ABs, batting averages and number of walks and the same number of extra base hits (so we can compare them equally) then I’ll take the guy that gets himself in scoring positon twice as often, even if he makes a few more mistakes on the basepaths. If the 25 SB player can get himself into scoring position with his bat significantly more often than the second guy, then the value the 50 SBs provide the second team is more pronounced.

It’s also possible that if the first player finds himself standing on second, third or watching his ball sail over the outfield wall a lot more often than the other guy he may have less steals because he has had less opportunity to do so as its harder to steal third or home than second. That’s not really the case in the Kinsler/Ellsbury comp, but it’s something that should be considered before automatically saying a generic player who steals 25 bases provides less value than one who steals 50 with a slightly higher CS rate.

by Curveball 2000 on Oct 9, 2008 3:13 PM EDT up reply actions 0 recs

Makes sense

and is more interesting than simply looking at the sb’s in a vacuum. A little disagreements though…

…While Kinsler is clearly a better player than Ellsburry at this point Jacoby ranked 12th among CF’s in Runs Created as a rookie and 13th in RC/27 outs and that is WITHOUT including his defense, which is clearly above average, perhaps one of the best. Very little of that offensive value comes from his base-stealing as well. He’s already an above avergae offensive player as a CF and with a little expected improvement he will be quite a player. Not even close to a bench player. He’s on his way to a Brett Butler type career IMO. Just like I predicted.

by casejud on Oct 11, 2008 3:26 PM EDT up reply actions 0 recs

Sorry that should read “project Kinsler’s numbers out over 145 games”. My bad.

by Curveball 2000 on Oct 8, 2008 2:37 PM EDT reply actions 0 recs

One, granted simple, interpretation

Linear weights suggest that Ellsbury was ever-so-slightly the more productive base stealer last year assuming I did the math right:

Ellsbury
(11CS * -.437) + (50SB * .193) = 4.843

Kinsler
(2CS * -.437) + (26SB * .193) = 4.144

I suppose this suggests that Ellsbury contributed 4.843 runs to the Sox while Kinsler contributed 4.144 runs to his team, but this is the SBs-in-a-vacuum analysis. I think you could use fangraphs.com to get the situations in order. One of the things you can find here is the win expectancy based on the situation but the thing that might help us in the Kinsler conversation is comparing WE before the play to after (as a result of the play).

It’s fun to tinker with, but I really can’t be trusted with anything more than pretty basic math…

by whonichol on Oct 8, 2008 5:22 PM EDT reply actions 0 recs

Just wondering

Did anyone else realize Matt Holliday went 28-2 SB-CS this year?

Who's world is it? It's yours.

by BlackOps on Oct 9, 2008 12:45 PM EDT reply actions 0 recs

If we looked at Win Probability Added

that would strongly argue for Ellsbury, given the bulk of his stolen bases. Each SB added roughly .016-.023, while his caught stealings went:

-.006
-.012
-.018
-.018
-.020
-.028
-.029
-.033
etc…

If we added up WPA on plays that were stolen bases or caught stealing Elllsbury comes out FAR ahead of Kinsler. He added much more to his teams chances of winning that Kinsler did on the basepaths. I think this is a method that fairly rewards a guy who attempts more steals.

by alskor on Oct 9, 2008 1:48 PM EDT reply actions 0 recs

LINK

fangraphs Ellsbury season play log:

http://www.fangraphs.com/statsp.aspx?playerid=4727&position=OF&season=&sort=

by alskor on Oct 9, 2008 1:49 PM EDT up reply actions 0 recs

Cool

This is basically what I have been putting off doing. I’d guess this is probably the best way to measure the more effective base stealer. Thanks alskor…

by whonichol on Oct 10, 2008 4:49 PM EDT up reply actions 0 recs