THE BOOK--Playing The Percentages In Baseball

Can’t this be interpreted another way? That the A’s and Rangers are wasting roster spots by carrying dead weight on their active roster when those players should be on the DL.

You can interpret it multiple ways, which is why you need more data.

By delaying the clock, they give themselves more chance to evaluate their player, rather than bringing up a minor league player (who may not even play anyway).

So, you can make arguments for both sides, which is why we need more data.

That data is really wrong. Maybe 90%. Not sure if that’s enough for most or the conclusions here, but using the MLB logs for this is a poor way of doing it.

Will/3

I wasn’t aware of any other source at the time. And trust me - were I aware of one, I would have used it as writing a script to parse MLB gibberish was a hassle.

Is there any freely available source of data that is more reliable than MLB logs? It would be great if you could point me to one.

Also, just to make it clear, the data in question is 2007-2009

Is there any study about how much a roster spot is worth? Or, in other words, how much poorer would an average team perform if they were limited to a 24 man roster?

Could one then make a connection what the negative value is for playing shorthanded a certain number of days, weeks...?

Good question.  Ok, so how much worse would a team be with a 24 man roster?  Worst-case is that your 24th guy does double-duty.  If you have a replacement-level reliever (win% = .470) and you use him up some more, maybe he’ll be a .420 reliever for the extra time?  Or even .370?  Let’s make him .370, or -.100 wins per 9 IP.  Given him 81 IP, and that’s close to 1 win.

So, that’s what I’d say.  The 25th spot is worth 1 win, if you intentionally don’t use it at all.

1 win on 180 days, or about .005 wins per day.

Two things jump out at me that will affect these numbers greatly.

1) The caliber of the player injured-If it’s your MVP CF you might be willing to hold off a few extra days whereas if it’s your 5th OF, he goes to the DL right away.

2) Starting Pitchers-If a starting pitcher gets injured, he’s not going to pitch again for at least five days and you are probably going to replace him with another starting pitcher who wouldn’t pitch during that time. Then you can always juggle your rotation around off days with little to no effect to give a few extra days rest/evaluation. And again, it depends on the caliber of the pitcher. Your #5 might get disabled right away while the ace of the staff gets a longer evaluation period.

Excellent points.  So, yes, we’d like to see the breakdown between guys in the starting rotation, starting lineup, bench and bullpen.  Good job.

One cynical view is that the retroactive dl move is yet another example of the a’s being frugal. You pay a guy his full major league salary when he’s on the dl and have bring up another player at 400k per year, about 50k a month. You delay the dl trip, and maybe the guy gets healthy enough to play, or at worst you call up a healthy guy 7-8 days later. In doing so you saved 8 days of service time and the pro rated league minimum for a week in the bigs.

Is that absurdly cheap? No more so than charging your players a dollar to use the soda machine in the locker room, or carrying a league-low 12 area scouts when they cost 30-50k apiece.

Tango, your second link to an elcroata article in one weekend, great stuff. And you probably found the second one by clicking on his archive and researching his previous work.

One of the things I love about this site is that you, with tons of knowledge already, are always willing to read new stuff and put in time. Your not ‘too big’ to do that. Awesome.

Dan/1, the alternative is that they put people on the DL immediately and end up in situations in which they’re replacing a major leaguer with a minor leaguer for say 10 days, when in fact they could have just given the major leaguer 5 days off and played with a roster of 24, and then had their major leaguer back for the final 10.

Bojan/4: No, sadly. I’d love to put mine out publicly, but even that isn’t perfect. The issue with the translogs is that it’s put together more from a press release sensibility than one of really monitoring roster spots. Ive talked with Tango about how “perfect” the data needs to be to find something useful. I certainly think this is a case where we can learn something, but I’m still unconvinced that the imperfect data doesn’t lead to imperfect results.

The other thing I’d suggest here - and I have NO idea how to do it - is seeing how many days are saved by holding off on putting a guy on the DL. Say a guy comes back in 8 days. That’s seven games “more” than if he’d been pushed to the DL on day 1. I think in a situation like what we saw with Carlos Gonzalez or with Ryan Braun, there’s a real value there. Given that the A’s and Rangers are among the smartest teams in terms of player valuation, both stat and scout perspectives, that they’re seeing some value in waiting.

Additionally, another possible conclusion one could draw from this is that it isn’t frugality on the part of the A’s or Rangers, but that the Yankees had more severe injuries, making the use of retroactive DL pointless. If a guy breaks an arm or tears an ACL, you’re likely not going to sit on a injury like that.

Will/12
That would indeed be valuable information.

The data would be even more difficult to gather, and unlike retroactive DL placement (however imperfect of a source MLB might be) not attainable in an automated way.

The only approach that I can think of is to manually track the day-to-day statuses of all the teams.

Bojan/14

That’s what I do. (Manually) There are HIPAA concerns about publishing that data as well, though I don’t believe it applies. I haven’t been able to convince lawyers.

Will:

My wife is a HIPAA expert, and she agrees with your lawyers that anything that would identify the specific player to a specific injury/medical condition would be covered under HIPAA.  Even though the teams have publicly released the data, you don’t have permission to do the same, unless you have a signed waiver from each player.

The chances of a player objecting to their medical info being released by you may be minimal, but if they did object, fines could apparently be substantial.

KJOK: so you are saying that the Hawks may be in violation by saying that Patrick Kane has a wrist injury:

http://blackhawks.nhl.com/club/page.htm?id=47391

Perhaps there’s a waiver in there between player and team.  But, if some third party were to post that same injury, they’d be in violation:

http://espn.go.com/chicago/nhl/story/_/id/6773324/chicago-blackhawks-patrick-kane-surgery-fractured-wrist

But, perhaps that’s a news story that’s being reported.  Is it instead that the NEWS is being COMPILED:
http://www.cbssports.com/nhl/injuries

As far as I’m concerned, and I’m no lawyer, once a player or team discloses information, then that information is part of the public domain.  Rights have been defacto waived.

Tell me why I’m wrong.

Professional teams almost certainly have received waivers from players to disclose their medical info, so no, Hawks would not be in violation.

But think about College Football - when a player is hurt during a game, teams no longer give the sideline reporters information about the injury, because if they did they would probably be in violation of HIPPA.  If they get player permission, they may give out that injury info after the game.

The ‘loophole’ that may allow CBS, or Will Carroll to disclose information that normally would be HIPPA protected would be that they are not considered ‘covered entities’ for HIPPA, which are generally Healthcare providers (which I think would include team medical staffs), Health care Clearinghouses, and Health Plans.  I’ll have to get more clarification from the wife, but I don’t know if CBS or anyone who compiles an injury database would be considered a ‘clearinghouse’.

KJOK: Exactly. The team has a waiver, but I do not. The problem isn’t in providing public info, but in “adding value” to that. A list would be fine, any conclusions drawn from that list or profiting from that list would not be fine. There have been a couple people that have put it out but I’m not willing to risk it against advice. Someone wants to buy my database and publish it, you know how to find me.

I’m not a doctor, so there’s the argument that I’m not covered, but again, I’m not going to be the one testing that legal argument in court.