rolling sabermetrics and statistics thread

https://www.reddit.com/r/baseball/comments/a747wg/hello_im_jonathan_judge_bachlaw_senior_metrics/

given the monumental wealth of statistical information available now, and advances in AI, surely 95+% of managerial decisions could be outsourced to a real-time computer? and if this is true, wouldn't it be borderline negligent not to?

some of them have been outsourced, to the degree that the field manager is expected to share the org's philosophy more specifically than ever.

Sabrites have said for awhile that allocation of playing time is more crucial than in-game tactics, and the front office has more to say about that too.

According to a friend who works for a team, most teams have their own proprietary algorithms and analytics departments. He likens SABR/FanGraphs intel to an open source resource like wikipedia.

nothing new here but https://www.nytimes.com/2019/02/13/sports/sean-forman-sports-reference.html

this is, hands down, my favorite statcast discovery so far

i looked at every two-pitch sequence (within PA but not across) in 2018 to see which combination is best at inducing swinging strikes

the result: throw the same exact pitch right before it! pic.twitter.com/8KZ8aaiGMM

— Alex Chamberlain (@DolphHauldhagen) February 21, 2019

i'd also be curious to see the analysis of the bad results of the throwing the same exact pitch twice. in other words, it leads to higher swinging strikes more often than other combinations, but maybe it also leads to a higher HR% or harder contact, etc - either it completely fools people or they're primed to wail on it. dunno, just guessing

yeah i can see going changeup-changeup being a high risk/high reward scenario

good point, karl

https://baseballsavant.mlb.com/affinity-pitchersAndHitters-byHittingProfile

explain that?

Enjoyed this:

Also wrote @ringer about how the rise of the opener is changing WAR, whether working as an opener or a bulk guy could affect a pitcher's earnings, and why unlike a lot of traditional stats, WAR never stays the same.https://t.co/cAswNEVH6c

— Ben Lindbergh (@BenLindbergh) March 25, 2019

Pre-move excavation...Found this list of the greatest hitters ever I drew up in 1976 or 1977, a couple of years after I bought my first Big Mac (and around the time I owned my first calculator). The very elegant formula I used is at the top. Keep in mind that a) I would have been 14 or 15, and b) I was still at least five years away from any awareness of Bill James or Thomas Boswell or anyone.

http://phildellio.tripod.com/hitters.jpg

The bad: 1) The randomness of the formula. There's an attempt to weigh the different elements (10 x HR vs. 1 x SB), but the final number (Ted Williams - 65.12) connects to nothing.

2) Strikeouts four times as important as walks! So Joe Sewell comes out ahead of Hank Aaron and Willie Mays. Um...I have no words of defense.

3) Runs and RBI front and center.

3) I'm not even sure the formula works. I remember finding this a couple of decades ago, and when I tried to calculate a couple of hitters, their overall rating didn't match what was on the paper. I suspect I calculated parts of the formula in some order that didn't follow correct order of operations.

The good: 1) An awareness that batting average and HR weren't everything.

2) Did I mention I was 14 or 15?

I'm very impressed with the printing in view of my illegible scrawl nowadays.

I don't know who the mystery player is at 51.05...

I suspect I calculated parts of the formula in some order that didn't follow correct order of operations.

You did! You just went left to right when there is a multiplication in the middle.

I don’t see strikeouts in the formula! Also, it penalizes walks I guess? Dammit 13 year old Clemens’s, wtf??

jk of course. This was a noble effort. esp knowing that you had to hand calculate everything. Weren’t you afraid you missed some players? (Unless...you didn’t do every player of all time...did you?)

Sorry, my phone autocorrected you to the rocket

Or wait, it’s adding for walks, and then subtracting 250

That's supposed to be 2 x SO - 1/2 x BB; strikeouts looks like 50. An ILX'or I'm Facebook friends with figured out how to make the numbers work out...Mike Trout comes in at 43-something, just below Mantle. The guy just strikes out too damn often, and I'm surprised the Angels don't cut him loose.

give me ed delahanty over any of these bozos

Weren’t you afraid you missed some players?

I also found pages of the final totals (not sure what to call it...CI, the Clemenza Index) for an alphabetized list of anybody who had a good career. I was working from the '74 edition of the Mac, not the first '69 edition; Aaron was still active, so I'm not sure what my cut-off date was (more likely, I had a minimum AB requirement).

Ed definitely makes the all-drinking team, from what I remember.

in 1895, delahanty scored 149 runs in 116 games. he hit a healthy .404 but came in second in the batting race behind jesse burkett at .409

it's a really cool list - awesome that you kept it!

despite my misreading of the equation, it is also very very neatly printed.

Would have generated a few posts on the ILB police blotter thread:

"Delahanty died when he was swept over Niagara Falls in early July 1903. He was apparently kicked off a train by the train's conductor for being drunk and disorderly. The conductor said Delahanty was brandishing a straight razor and threatening passengers after he consumed five whiskies. After being kicked off the train, Delahanty started his way across the International Railway Bridge connecting Buffalo, New York with Fort Erie (near Niagara Falls) and fell or jumped off the bridge (some accounts say Ed was yelling about death that night). Whether "Big Ed" died from his plunge over the Falls or drowned on the way to the Falls is uncertain. His body was found at the bottom of Niagara Falls two weeks after his death."

don't ask why i felt compelled to make this, but might as well share. this is the top 5 wRC+ for each year of the 1950's. i wanted to see which player seasons were way ahead of their peers, relatively. williams in '54 and williams and mantle in '57 both stand out

https://i.imgur.com/5tSykHj.png

also, i have to share this as well because it is crazy how the charts in Excel wannabe clone Numbers are hard to customize and offers all the wrong options...unless you go into 3D graph mode, in which case everything is possible

https://i.imgur.com/Sr0KzWP.png

Surprised there's such a difference between Mantle's '56 and '57--his bWAR is the same for the two seasons. I know it's relative to the league, but still.

heh, the chart i made is confusing as all hell. it's so confusing even i can't figure out wtf is going on!

but basically, the top line represents the first place finisher in wRC+ for the year. it happens to be williams in both '54 and '57, which makes it look like the entire top line is williams. but in '55, for example, the first place wRC+ was mantle, at 179.

...i know, that's really confusing! i only did it that way because i wanted to see the discrepancy between the top 5 finishers each year, to look for outliers.

mantle's '56 was actually very close to '57, by wRC+. it was 202 in 1956 (twice as good as the average hitter!!), which was first place in the league, and 217 in 1957, second only to williams.

*fingers crossed behind back*

i will never make another ill-advised graph again

C'mon, the 3D one even casts a shadow--worth the price of admission alone.

not exactly sabermetrics, but: the details behind pitch classification

https://technology.mlblogs.com/mlb-pitch-classification-64a1e32ee079

this seems like a wild stab in the dark to me, but

We created a dashboard for NCAA WAR over the past four seasons. Calculate WAR based on school, year, name, and more: https://t.co/ZI6q74BApb

You can read an introduction to cWAR and what went into the dashboard here: https://t.co/PvFB5KGiKV pic.twitter.com/ftkYyfEsW1

— Driveline Baseball (@DrivelineBB) June 10, 2020

this is great

https://www.theringer.com/mlb/2020/7/17/21328094/analytics-60-game-season-war-stats

i'm with John Thorn

Dallas Keuchel is 6-2 w/ a 2.19 ERA. He has 1.7 bWAR & 1.5 fWAR.

Rick Porcello is 1-5 with a 6.06 ERA. He has 0.0 bWAR.& 1.5 fWAR.

— David Laurila (@DavidLaurilaQA) September 17, 2020

thing is, both bWAR and fWAR are useful, for pitchers, they just measure different things. bWAR lines up with what people think they're getting out of WAR, and is based on results. fWAR for pitchers is based on FIP, so it's more of a measure of how they "should" be doing. which is confusing, agreed on that. wish that both b-r and fangraphs would offer both versions - give me both a results-based WAR and a peripheral based WAR, and just keep them separate imo

i like having them both because i can look at porcello's line and immediately know that he's been unlucky in terms of his results (likely he has been giving up a ton of home runs, or his LOB% is way lower than it normally would be because of bad sequencing luck).

actually porcello's given up very few homers and his babip's a nightmare, which is no doubt why fangraphs likes him

i'm willing to grant that he's been unlucky with balls in play and LOB (the mets defense probably doesn't help). but i suspect his 5.7% HR/FB this season (12% for career) is luck too. meanwhile his WHIP is over 1.5, he doesn't have great strikeout numbers for this day and age, and he goes less than five innings per start.

and keuchel's been *insanely* lucky in every facet! so i guess i feel like they should both have lower fWAR

(this is not to argue with your larger point, which is otm)

it also seems likely that i have an outdated idea in my mind of what one should expect from a (good/average/replacement-level) major league starter

Starters? pfffft. Give me a listicle of the top 500 openers in the game (globally, including little league) and I'll be projecting HOF eligibility by sundown.

there's always been this...WAR (sorry)... between BR and fangraphs with the WAR stats. honestly, i think fangraphs needs to change, because their WAR philosophy inconsistent between the pitchers and hitters. for hitters it seems more (though not entirely) results-based, like BR. but for pitchers it's more peripheral. it's just kind of weird, especially when you're trying to compare value between pitchers and hitters.

Do you really want to get me started about how flawed their defensive metrics* are and how that number is quietly folded into each site's "offensive" WAR and just published as QED gospel?

*just about every other season there is some radical makeover of defensive stats causing some huge ripple in the lists that appeal to whatever you call the devotees of "MOJO-magazine equivalents of baseball"-type content.

oh yeah, i feel you on that for sure. i've never understood why it's not just clearly bifurcated as offense WAR and defense WAR.

it is! At least on bbref it is, there's oWAR and dWAR (which doesn't actually add up to WAR because it includes the position adjustment bonus in each)

oh d'doh, i just mean on fangraphs. i still hate the baseball-ref style/format, and rarely visit it for that reason alone. you know how there are chrome extensions? and particularly for some websites (like discogs) that reorganize the data on the fly? i wish someone would do that for either baseball-ref or fangraphs.

Good Posnanski piece that definitely spotlights a major flaw in bWAR.

https://joeposnanski.substack.com/p/burnes-baby-burnes?utm_campaign=post&utm_medium=email&utm_source=copy

(That's his "share this piece" link, so should work.)

the fog of glove again

great article, thanks for taking the time to share!

however, i ran across a similar anecdote the other day and removed it from my memories, and i am afraid i am about t delete this again for similar reasons:

*By the way, do you know who IS playing in front of the 1970s Baltimore Orioles defense? Adam Wainwright. DRS does show the Cardinals being a terrific defensive team, but even that underrates how good they’ve been behind Wainwright; the Cards are 22 Outs Above Average when Waino is on the mound. That is more than double anyone else in baseball.p

i don't know what the previous series of deleted words was about, but one thing i know is that adam wainwright is on his way to be being the best pitcher in baseball again, and it's definitely related to luck!

i prefer fWAR to bWAR a lot for pitching, though i find that bWAR defenders tend to be way more defensive about it than fWAR defenders. people really just don't like the concept of FIP. but the thing tango said about halving the value of DRS should apply to hitters too... it results in massive outlier seasons because DRS is so springy when UZR is more conservative. it really seems to impact MVP conversations too even though no one really trusts defensive stats. you get these wild 10+ WAR hitter seasons because someone put up a DRS of 8 million (and a UZR of 15)

should also point out that fWAR isn't just FIP, there are also adjustments for IFFBs and catcher framing (though that's controversial for some people too)

that said i do appreciate that bWAR updates its position adjustment values season by season, fangraphs just keeps pretending that playing shortstop is as valuable now as it was 20 years ago despite teams now feeling comfortable parking their trucks there

yeah it is a fluid situation for sure