Discussion Question: Least Favorite Stathead Stat

I thought this said “least favorite stathead” and left it at that. Could of had an f’ing field day with that one.

My own opinion: I don’t know why, but the obsession with BABIP gets on my nerves sometimes, even though I use it myself.

I 100% agree. I think people overuse this stat and if a player is struggling they automatically turn to it and say it’s the cause. If a player is struggling he might not be making solid contact, which can lead to a lower BABIP. BABIP isn’t all luck, it’s caused by multiple things.

I’m interested to see how the next frontier-PitchFX Data-effects the way we look at batted ball data. I have a feeling we’ll go beyond line drives and start talking about velocity off the bat, area-of-field selectivity, and all sorts of things I can’t conceive of.
Exciting stuff.

cause that is the one that bothers me the most. I think it’s a great stat, but I think too many quote it as gospel. They try to paint every player with the OPS brush and don’t account for different types of players. Lead-off/speed types of players are OPS outliers I believe.

But of the newer stats that I don’t much care for, I’d have to go with FIP. I like to check out all the stats a pitcher puts together, and ERA is one of them. FIP just takes into account homers, walks, strikeouts, etc. I can see all that. But it doesn’t tell me how a pitcher does in a jam. Give me two pitchers with similar FIPs and I will want to see an ERA. I want to know if one of them is allowing less runs, because it could be indicative of how they do with runners on, pitching out of the stretch, etc. I like old-fashioned ERA, and then I can look at the component stats to see if it was a bit of an illusion or not.

use OPS when discussing players like Bret Gardner, Michael Brantley, Jacoby Ellsbury, Nyger Morgan, Elvis Andrus, etc. Most of these speed types of players will have a sub-800 OPS.

for the one reason that there is so much variance between different stats comparing the same players. Whenever there is a discussion about defense, there are inevitably the couple of posters who will pick and choose which stat fits their viewpoint and then act like that is infallible, while another defensive metric they often use says something else.

For some reason this one bugs me. It is just a linear weights rate stat artificially re-jiggered to look like OBP. So, if someone says Player X has a wOBA of .400 this does not tell me much other than I know that a .400 OBP is good so this must be good, I guess.

I have the same problem with Clay Davenport’s eqAv. This does not do much for me for some reason. I much prefer the old style (as in Bill James) measures of RC/27 outs. I just find this more meaningful to me personally.

I do understand there some advantages to the Tango (wOBA) and Davenport (eqAv) derived stats, but I have been able to fully embrace them.

oops wrong site

they all seem to have a long ways to go before they are dependable means of calculation

Not so much WAR in and of itself, but forgetting that there are a lot of moving parts causing a lack of precision—especially when it is taken as gospel as to whether a player has over/underperformed his salary. At the very least, it is often forgotten that the precision with WAR is generally overstated — the bigger picture can be seen, but the details really can’t.

I mean, it’s fun to graph out a game, but some people try using it like it actually means something. It’s a junk stat, nothing more.

I also hate OPS but like you said that doesnt count

“WAR” is not a statistic. Wins Above Replacement is just a model, built from a set of assumptions.

There are many, many different types of WAR, they all define the parameters differently. For instance, Fangraphs’ WAR uses wOBA as the offensive metric and UZR as the defensive metric. Rally uses Total Zone. Etc..

If you don’t like the way somebody has defined one of the parameters, that’s OK. Do something else. If you prefer +/- to UZR (I do), use +/-. If you’d like to build your metric with WPA/LI instead of context neutral linear weights, do that. The model, however (for pos. players, batting + defense + positional + replacement, for pitchers, (((((LgRAA/9) – (RAA/9))/run environment ) + .500 – Rep) * IP/9), is valid. If there’s a good argument this model doesn’t work, I haven’t seen it.

If you’re going to say “WAR”, you need to specify which type of WAR and which parameter you disagree with.

is useless

I think we still have a long way to go with defensive metrics