Petey's New Stat Corner: ED'S OPS

Bruce, Stumps, Frazier – Laynce

Is that a Trapper Keeper cover?

It is not adjusted for park or league, which are both significant. The difference in OPS between the leagues is over 13 points of OPS during the last decade.
Doesn’t the presence of a DH instead of a pitcher account for the 13 OPS point differential between leagues?

Because your data doesn’t account for league, we have no way of knowing. It could be the DH, better hitters overall, or better hitting environments (parks). Since you control for none of these, you’re just guessing.

while the NL OPS is .752 over your study period, it has varied from .739 (just last year, incidently) to .773; the AL has varied from .755 to .792, yet the average is .765. To connect with my first point, the difference between the leagues varies even more year-to-year: As little as 4 points in ’07, as much as 25 points in ’09.
It is BECAUSE of the fluctuations you mention that I have selected a 10-year sample here to include more examples from which to draw information, and to help dilute any outliers, positive or negative.

Just in the NL, eight teams (half of the league) play in different ballparks than in 2000; the Nationals have played in 3 different stadiums during the decade being studied. Outstripping all of those, though, is Colorado: thanks to the humidor, Coors Field has gone from being a +25% run environment to +10%. Yet, your numbers have no way of reflecting any of these changes other than to assume they all just “even out”.

You have set a strict limit of 502 PA, yet lower that for catchers to get 20 (19.3, actually) starters per year. Either 502 is your standard, or 400 is. If not enough catchers qualify, then they don’t.
You’re welcome to your opinion, but I disagree. Even the most hearty of catchers are routinely granted a day off when a day game follows a night game. Also, catchers seldom benefit from the additional plate appearances afforded in pinch-hit opportunities or within double-switches as managers traditionally prefer to keep a catcher ready on the bench in chance of injury. A catcher who gets you 400 or more plate appearances in a year is a starting catcher, IMO.

I appreciate that you put alot of work into this, but the fact that you completely ignore context makes this stat as useful as raw OPS, and your baseline is so high as to make an All-Star “average”. I ran your criteria through the Play Index at baseball-reference.com for a couple of positions (I used 3000 minimum PA, 50% of games played at the position). Do you know how many players qualified as an “above average” 1B over the decade? Ten. Second base? 9. LF? 8. To me, that makes this a useless stat as it is.

If you wanted to do a study to use for the 2010 offseason, perhaps a 2006-2009 study would be most prudent…but it will still include some “Manny bring Manny.”

Fuckin’ ’roiders ruining our stats…asswipes!

Can you point me to a reference?

I could see how large samples might get more outliers, but the fact that small samples have large errors in the mean is fairly rigorously proven.

Once that leaves him, he’s pretty much replacement level in my book due to his declining bat.

In the proper context, they can offer some valuable insight, but unfortunately I just don’t think can be hailed as the gospel due to the fact all of them are still works in progress.

UZR is probably the most “widely accepted” defensive stat right now, but even it’s not perfect. Defensive WinShares is also in its infancy, IMO. I eagerly look forward to the day when they both follow a consistent and testable methodology. It seems like every other year the defensive stats tweak and re-examine their calculations/equations to improve, which is a great thing, but with evolving change I think it’s intellectually irresponsible to accept their results as absolute truth at this point.

Speaking of point, what is my point? LF is the one position where I’ll fight to my death screaming that defense simply does not matter for the player, but after watching Griffey lose it in CF and exiled to RF, and the butchers the Reds trot out to SS year after year, and the spectacularness of Brandon Phillips, I easily admit that defense does have a legit claim as part of a player’s worth…. the only problem is that we cannot accurately quantify that defensive value, yet.

the Fan Scouting Reports with UZR does a serviceable job. I’m pretty onboard with anything that takes advantage of emerging intelligence and takes a consensus of multiple metrics — sort of the Nate Silver + Malcolm Gladwell approach. I like that it compensates for a player being over/under-valued by the system in part by directly reflecting how the player is valued by fans, which might actually help to neutralize the effect since fan appeal translates to economic value. Teams do have to weigh who they keep on their team based on whether fans will turn out to support them – witness Edwin and Dunn. And their instincts might be pretty on the mark if you consider just how bad Dunn’s defense has been the past couple seasons.

This stat is chiefly for comparison between everyday players whom you might acquire through trade or FA signing, and comparison between your existing options.

I just don’t see why you need it. When you are looking to sign a player, you generally know what his role will be, and you don’t need a stat like this to find comps for a player. Besides, if you want to compare to individual players, why not just use their individual stats?

If you are looking to sign Corey Hart (for example) to be your LF, you know he isn’t going to be average with the bat compared to full-time LFs in the league, but that isn’t really the question at hand. It doesn’t make sense to put him in a comp with Holliday, Dunn, Braun, etc. The question is whether he is better than the players that are the options for your team, which is a much lower baseline. Is he better than Laynce Nix or Wlad Ballentien? Is he better than Jason Bay, considering the money to be spent; that is, only if your team could afford to sign Bay as a FA in the first place. If not, there is no reason to include Bay in your comparison at all, as he is not an option and has no bearing on the decision.

This stat might be useful in contract negotiations. Say a player wants to be paid like one of the top LF in the game. Then you see how he stacks up to the players covered by this stat, and see how the FAs in that group are paid. Still, it’s easier to simply pull individual comps (which is what agents and teams typically do) than to include an entire sample of dissimilar players.

My basic complaint is that this stat does not tell you anything that you don’t already know without the stat. To put it more clearly, it doesn’t add any information that would affect a personnel decision. it doesn’t evaluate players, it simply compares them to an arbitrary standard. How would it influence a roster decision?

i might finally coax the lil’ lady to go tomorry though.

I don’t disagree with your player percentage concept, although it invokes a troubling amount of personal judgment per situation that I’d like to avoid. I prefer a formulaic approach where an established standard is applied uniformly.