NERD for Position Players, Version Two
Last week, I submitted for the readership’s consideration a first attempt at NERD scores for position players. In this post, I submit a second iteration with more variables, as suggested by a number of readers.
For those unfamiliar with NERD, allow me to explain briefly. NERD is an attempt to issue a score (0-10) to, in this case, every position player with more than 100 PA, on the basis of how interesting or watchable that player might be to the learned baseball fan.
While it’s obviously impossible to anticipate every person’s taste — and equally impossible to find in the numbers whatever brand of je ne sais quoi Nyjer Morgan possesses — it’s also probably the case that saber-oriented fans have enough in common to know that Jose Bautista is excellent and Willie Harris, while assuredly affable, does little on the field to merit attention. The benefit of NERD is that, like other stats, it “watches” all the games, so it can help us identify and enjoy players we mightn’t have otherwise known we liked.
For the first iteration of Player NERD, I used only three inputs — namely, age, WAR (per 650 PA), and BABIP luck (where being unlucky is worth a positive value). A number of readers suggested that, while that constituted a decent start, there were other things to consider, probably.
For the present iteration of Player NERD, I’ve added an additional three inputs that I feel address the (helpful) advice of those commenters. Those three variable are home-run per fly-ball percentage (HR/FB), Speed Score (Spd), and strikeout rate (where lower equals better). These variables account for three things which certainly affect the aesthetic value of a player: power, speed, and the ability to, you know, make contact.
Here are the top-20 players now using that method (were N2 = Nerd v2, N1 = Nerd v1, and Diff is the difference between the two scores).
Here are the 10 players who benefit most by the new scoring system:
And the 10 who fall farthest:
While the inclusion of Willie Bloomquist in the “risers” table is perhaps a red flag, there are a number of players on that lists — Albert Pujols, Curtis Granderson, Ryan Braun — who probably do offer something besides what merely their respective ages and WAR might otherwise tell us. Conversely, Clint Barmes, while assuredly affable, is not particularly entertaining.
Two last things, then:
1. See the spreadsheet of all the 302 qualified players here.
2. Thoughts?
Carson Cistulli has published a book of aphorisms called Spirited Ejaculations of a New Enthusiast.
I think HR/FB and Speed Score are good additions. However, as a TTO fan, I can’t approve of K% being used in this fashion.
But, despite the fact that strikeouts are little different than an out, they are more frustrating, and make people hate guys like Adam Dunn without true warrant.
Oh, definitely the arm rating from UZR.
Seconded!
yeah, probably a good point
My only thought is that if the WAR being used includes a positional adjustment it could skew the results slightly due to the positional adjustment rather than the player’s watchability. If so, taking the positional adjustment out of the WAR formula might yield a better result (or alternatively, breaking the WAR component into separate offensive and defensive metrics, e.g., wOBA and UZR).
Obviously this is a question of taste, but I’ll submit that, in my head, I’m making mental adjustments anyway for position. Like, if I guy can hit but I KNOW he’s a poor fielder — or, if not poor, then an average 1B or LF or something else on that side of the spectrum — then I’m adjusting for that anyway. He NEEDS to hit.
Like, last year, when Adam Lind wasn’t hitting — or this one, with Adam Dunn (who, I recognize, is awesome for other reasons): if those guys aren’t hitting well, it’s even more dissatisfying, because that’s why they’re on the roster.
The WAR positional adjustment just makes that explicit.
I have no objections to the formula, in general, anything amusing is good, particularly, attempts to quantify awesomeness. My concern, if you can call it that, is marginal catchers/middle IF getting higher scores than more watchable 1B/corner OF due to positional adjustments inside the WAR formula.
For example, Jay Bruce and Danny Espinosa are tied in Fangraphs WAR (my guess is due to positional adjustment), but I think many “of the sabermetric persuasion” would find Espinosa less watchable than Bruce. The objective of the player NERD score (as I understand it) is to reflect this difference in watchability, however, they are tied in player NERD score.
This will keep me up at night, once I finish splitting a few more hairs.
But Danny Espinosa is a stats nerd’s God.
I don’t know how seriously one can take this if Miguel Cabrera isn’t even rated…
He rated 55th with a Nerd of 7. It’s all in the spreadsheet.
Cabrera is 54th, right between Mitch Moreland and Mark Teixeira. He’s hurt a bit by the BABIP formula, which is biased against players like Cabrera who always have a high BABIP regardless of luck. Not sure how to adjust against that. Maybe needs to be a comparison to career BABIP.
He’s also being penalized for a low speed score, which in my mind presents too much of an extreme between the high and low for something like this. A player that is faster than normal is going to be more exciting to watch, but I don’t know that someone with a below average speed score becomes any LESS interesting to watch, I don’t think any player is slow enough that their slowness is noteworthy.
Maybe plus points for a positive score, with no bonus for a zero or below is more accurate?
Interesting point about Spd there. On my phone now, but will run a version like that in a bit.
He needs to use the difference between BABIP and xBABIP for each player.
Well, perhaps it’s a little more complex than that. True, a slow player like Cabrera more than makes up for his lack of speed by being a very threatening bat. Can the opposite type of player be as interesting? That is, can someone who’s extremely quick but little power be interesting, such as Ichiro?
I think its the extremes that interest us more. We want extreme power or extreme speed. Being merely good at both isn’t quite as interesting.
I would love to see IsoD included.
god knows nothing’s as entertaining to watch as someone drawing a walk….
To me, this is getting much closer. The top and bottom pass my initial sniff test, which is if I was walking in the mall past a shop selling TVs, would I stop to watch the at bat? Just glancing at the top 20, for many of these guys, the answer is yes, I would.
I assume that you are going to adjust the overall game NERD for those games featuring a large number of high or low NERD players? Are you going to weight the extreme 10s and 1s?
Hey Carson – looks great. Intuitively I like the v2 better. Seeing Asdrubal on the top got me wondering if you could add anything for breakout year players? Or maybe even streaking players like Ethier earlier in the year?
Kind of like Hollinger said about speed, I am interested to watch someone playing above their head (or breaking out), but less so to watch somebody keep struggling through a slump so maybe a bonus for streaking but no penalty otherwise?
The problem with streaking is it’s tougher to quanitfy what makes a streak. It can stop and start so quickly that it’s almost not worth the effort. However you could try to measure it by giving a small bonus based on Positive Change in NERD score over the last 7 days, or something similar. That could capture the moment in time– this player need to be watched RIGHT NOW, but maybe not otherwise– effect.
BACON! Seriously, check it out:
http://www.plunkeveryone.com/?p=95
Also, what about a small component for HBP? That’s always kind of fun.
When miggy is the third best cabrera, not first, you are missing something. I don’t care how good the other ones are doing he is still the best and most dangerous hitter in baseball.
I wish there was some way to quantify “prettiness of swing,” which is a lot of the reason I love AGon
Maybe include how often they’re tossed from the game/argue with umps? Is there even a stat for that?
NERD was a featured statistic on a Mental Floss quiz posted recently!
http://www.mentalfloss.com/blogs/archives/89769
Outfield arm component from UZR definitely makes a player more watchable. Every time you see a Francouer game, you pray that the 3rd base coach sends the runner, or someone tries to tag and advance on a fly ball, just to watch Francouer unleash the cannon that he had his arm replaced with. This also explains why he’s so terrible with a bat.
I think a bonus should be given for walk rates and pitches per plate appearance, because long drawn-out at-bats where a player gets down in the count, fouls off pitch after pitch, then manages to still work a walk 9 pitches later are still entertaining.
One question I might ask is whether this statistic is trying to tell a story of what happened or what will happen? If it’s what we think will happen we could to re-create WAR using larger sample sizes for the numbers that take longer to stabilize — here we’d have a sort of ZiPS like NERD for Position Players bumping down players like Bloomquist who have been, surprisingly, very watchable this year but perhaps won’t continue to be while bumping up players like Choo or Ichiro who have been hard to watch so far but probably will be more watchable going forward — if it’s what has happened the xBABIP is a tricky addition. Not only is xBABIP a little controversial, if I remember correctly, but it’s certainly a predictive statistic. I suppose NERD for Position Players could try to tell both at the same time, i.e. what will happen based off of this year’s data, but it’s not the only way and for trying to understand things like a players defensive contributions we might be better off looking backwards a little farther?
Secondly, are there any plans to eventually list this statistic somewhere, or will we have to just catch glimpses through the One Night Only feature?
Longoria went from first to outside the top 20? wow
This needs a component to account for number of times an outfielder throws his glove in anger after missing a catch, number of times a player attempts to clothesline a pitcher ~8″ taller than himself, and number of times a player railroads the catcher. Oh, and triples. Lots of triples.
I’m not sure that age is more important than experience. If a player debuted at 21 and has been doing well enough to still have a job 2 years later he is probably not as interesting to check out as the other 23 YO who is a rookie playing as well. But I would agree there is some curiosity value to checking out the 19 YO HS-drafted rookie rather than the college-drafted 23 YO rookie in case it is your first glimpse of the next 300 game winner.
You have a suprisingly narrow view of what it means to be a “learned baseball fan.” It seems to exclude a lot of learnedness.