Sports Analytics ... what are your thoughts?

[Tinwoodsman]

Is the train off of the tracks?

[Nova Carmina]

6 hours ago, Tinwoodsman said:

Is the train off of the tracks?

Yes.

 

Generally, I like analytics, and I think it has improved the ways we can talk about the games, but at some point it seems like the theory has separated from the reality.

 

There's an old corny joke I love about three statisticians who go duck hunting: after the first one misses high, the second one misses low, and the third one shouts "We got it!"  

 

But you didn't, you see.

 

Because my primary baseball team, the KC Royals, are really awful, the broadcast team are constantly having to look for silver linings, and one of things they keep harping on is "hard hit rate." So-and-so is now 0-23, but man that last line drive right at the shortstop was really sizzling. When I was in high school, my coach -- like a million coaches around the country -- told us to "hit it where they ain't" (is that a Yogi-ism?), and it didn't matter if it was 104 off the bat or 64, if it landed on the grass somewhere, it was good, but now the measures seem to be the point, rather than the result, and so we get defensive shifts that look like power lines when a hawk is at one end and all the other birds are way the hell at the other, we get "three true outcomes" baseball with home runs and strikeouts galore, we get writers like Bill Barnwell who insist the process is more important than the result, like real games are settled by mathematical formulae, like it's D&D or something.

 

I blame fantasy sports gambling for putting such an emphasis on stats and less on winning and losing.

 

 

[Tinwoodsman]

40 minutes ago, Nova Carmina said:

Yes.

 

Generally, I like analytics, and I think it has improved the ways we can talk about the games, but at some point it seems like the theory has separated from the reality.

 

There's an old corny joke I love about three statisticians who go duck hunting: after the first one misses high, the second one misses low, and the third one shouts "We got it!"  

 

But you didn't, you see.

 

Because my primary baseball team, the KC Royals, are really awful, the broadcast team are constantly having to look for silver linings, and one of things they keep harping on is "hard hit rate." So-and-so is now 0-23, but man that last line drive right at the shortstop was really sizzling. When I was in high school, my coach -- like a million coaches around the country -- told us to "hit it where they ain't" (is that a Yogi-ism?), and it didn't matter if it was 104 off the bat or 64, if it landed on the grass somewhere, it was good, but now the measures seem to be the point, rather than the result, and so we get defensive shifts that look like power lines when a hawk is at one end and all the other birds are way the hell at the other, we get "three true outcomes" baseball with home runs and strikeouts galore, we get writers like Bill Barnwell who insist the process is more important than the result, like real games are settled by mathematical formulae, like it's D&D or something.

 

I blame fantasy sports gambling for putting such an emphasis on stats and less on winning and losing.

 

 

Excellent reply! Thanks!

[laughedatbytime]

I want to write up a defense of analytics (or what mostly will be) but I don't have enough time to finish on my lunch hour.   Be back later for what I hope is a fun discussion.

[blueschica]

I liked baseball better before analytics became the end all to be all.  There is no subtlety left.  No walks, no bunts.  Pitchers being pulled IMO prematurely.  Hopefully things will swing back to a more middle ground someday.

[Tinwoodsman]

10 hours ago, blueschica said:

I liked baseball better before analytics became the end all to be all.  There is no subtlety left.  No walks, no bunts.  Pitchers being pulled IMO prematurely.  Hopefully things will swing back to a more middle ground someday.

Yes, baseball has become very different and not in a good way.

[Tinwoodsman]

12 hours ago, laughedatbytime said:

I want to write up a defense of analytics (or what mostly will be) but I don't have enough time to finish on my lunch hour.   Be back later for what I hope is a fun discussion.

No pressure! I look forward to your comments!

[Tinwoodsman]

One thing that I did not expect to see in the wide wide world of sports is a resurgence in violence. I'm not sure if there is a correlation between analytics and the work environment or if our world is in the process of a reset due to the residue from covid. Monetary fallout  hasn't helped. I see it where I live too ... there are some wildcards out there!

[Chicken hawk]

Poised Celtics start fast, hold off Warriors for 2-1 lead in NBA Finals

[pjbear05]

On 6/9/2022 at 2:10 PM, Chicken hawk said:

Poised Celtics start fast, hold off Warriors for 2-1 lead in NBA Finals

Curry drops 40 on Celtics, Warriors tie series 2-2 with 106-97 win.

[laughedatbytime]

On 6/9/2022 at 12:21 AM, Tinwoodsman said:

No pressure! I look forward to your comments!

OK here goes...

 

In defense of analytics...

 

Analytics at its core is just gathering data, asking questions, and using the data to answer the!m.   This can, like anything else be done well, and also done poorly, but teams who do it well have advantages over teams that don't.

 

There are ways in which analytics have indelibly changed the game in terms of how the games are played (the effect on aesthetics can be negative, but that's a different debate.)   Two of the most obvious are the shift in baseball and the shot distribution in basketball, which is more heavily weighted toward dunks and threes at the expense of the mid range game.    These are pretty clearly successful strategies and teams that don't employ them are at a distinct disadvantage.

 

Some of the more recent innovations in data collection have allowed analysis of launch angle and exit velocity and spin rate for pitchers have also altered the way the game has played.    A focus on launch angle has allowed formerly mediocre players like Max Muncy to become valuable to teams.  Exit velocity is one component of analyzing hitters and can help determine which hitter is better in a more thorough way than looking at traditional stats like batting average.   xwOBA is far more predictive than batting average.

 

Jokes about statisticians aside, what these newest stats and measurements do is allow for improved best estimates of the prediction of future performance of players and helps identify areas which can be worked on to improve expected player performance.  No statistician worthy of the name would claim that they knew everything or that anything is perfectly predictive, only that it should improve expectations for winning and better performance given enough time for the randomness of outcomes to work itself out.    Billy Beans said it best when he said something like "this shit doesn't work in the playoffs" by which he meant small samples lead to high variance of outcomes.   It's not a condemnation of analytics, it's a recognition of their limitations...the relatively tiny but positive impact on outcomes is overwhelmed by the inherent randomness of events.  Appropriate use of analytics help teams win, but it isn't the sole reason; anyone making the claim that it does should rightly be mocked, but setting up that straw man and knocking it down shouldn't be considered to have shown that analytics doesn't matter.

 

Addressing the effect on the aesthetics of the game, I don't think its arguable that it's hurt, in baseball and basketball, at least; the shift and the three or dunk shot selection have made the game less aesthetically pleasing.   If this is a concern to those who run the leagues, it's incumbent on creating incentives for winning that match that aesthetic; you play to win the game, and you develop strategies to optimize that outcome.

 

I'm out of time for now, I did want to address (real) football analytics but there's enough differences between it and baseball and basketball that it deserves its own post.

[Rick N. Backer]

11 hours ago, pjbear05 said:

Curry drops 40 on Celtics, Warriors tie series 2-2 with 106-97 win.

That was a HUGE loss IMO.

[Nova Carmina]

Management guru Peter Drucker: “What gets measured gets managed – even when it's pointless to measure and manage it, and even if it harms the purpose of the organization to do so.” Accounting historian H. Thomas Johnson: “Perhaps what you measure is what you get. More likely, what you measure is all you'll get.

[laughedatbytime]

So better measuring (and the correct attribution of certain traits to success) leads to better results?   Sounds about right.

 

This key isn't the data, the key is determining what contributes to winning and what doesn't and then determining how to use  it to build your team.  Ignoring the latter and emphasizing the former should lead to success.   Analytics are A tool, not the only tool, but one that should not be ignored.

[edhunter]

I get why they're used. I recently watched Moneyball and really liked it. But it makes the game unbearable to watch. Hopefully things like pitch clocks, eliminating the shift and cutting down on pitching changes will counteract them.  They're effective over a long sample size, but in a short series, you still need to rely on baseball instincts. Tampa blew the World Series a few years ago because of over-reliance on them. 

 

 Another potential downside to them is this: My behated  Red Sox are owned by John Henry. He views the team strictly as a part of his sports business empire. He also owns the Pittsburgh Penguins and Liverpool FC. They use analytics to ensure a certain degree of success, and profit.   Last year the Red Sox made the American league championship series. Before the trading deadline they ran over 50 analytical models which showed that no combination of players on the market would have gotten them to the World Series and won, so they declined to make the investment in them. They passed on Max Scherzer, and Anthony Rizzo and settled for Kyle Schwarber.  They used analytics to choose maximum profitability over a chance for a  championship.  The revenue of 2 more home playoff games and the prestige  of a potential championship was not worth the risk of the extra investment to secure them.   It may have been their last best chance, because several of their best players are free agents this year,  and it looks like there's no way they're going to spend the money to keep all of them.  They have taken the focus away from the game and put it on the profits derived from the game. The results of the game are irrelevant  past a certain profit point.

Edited June 16, 2022 by edhunter

[grep]

I'm not math guru. But I imagine that the law of averages eventually catches up with you.  You have X likelyhood of doing Y in Z situation..... so if you start acting on that the same way all the time, eventually the stat becomes meaningless....or self re-inforcing depending how yuo want to look at it.

it's like, behave to the stats and the stats become the game. Rather than having at it, thinking outside the box, and doing something new.    Seems pretty boring and generic, if you ask me.

Yeah I know, variations in weather, does the pitcher sneeze.....are supposed to keep it interesting. But does it really? Or is all that already baked ito the stats?

[edhunter]

14 hours ago, grep said:

I'm not math guru. But I imagine that the law of averages eventually catches up with you.  You have X likelyhood of doing Y in Z situation..... so if you start acting on that the same way all the time, eventually the stat becomes meaningless....or self re-inforcing depending how yuo want to look at it.

it's like, behave to the stats and the stats become the game. Rather than having at it, thinking outside the box, and doing something new.    Seems pretty boring and generic, if you ask me.

Yeah I know, variations in weather, does the pitcher sneeze.....are supposed to keep it interesting. But does it really? Or is all that already baked ito the stats?

Now that virtually every team is using them, the small market teams are basically back where they started. There's an Ivy League nerd team in every front office, so the hidden gems like Scott Hattiesburg (from Moneyball) are harder to find. And if they do happen to find one, a bigger market team will outbid them in free agency.

[Nova Carmina]

Maybe there was never a time when every time was trying to win the whole thing, but modern ownership has often used analytics -- as in Edhunter's example with the BoSox above -- as an excuse/reason not to try to win: "We ran the numbers, we can't do it, so why bother?" They want the teams to be profitable first, and if that involves winning, so be it. It would be nice if those things were more closely correlated, but the Cowboys are one of the most valuable franchises in the world, and they haven't won shit in nearly 30 years.

 

As LABT explained upthread, the massive amounts of data should be employed to enhance the likelihood of winning, and sometimes that might mean winning "ugly" (his point about the aesthetics of the shift, for instance), but if every team is shifting, and some teams (Let's go, Royals!) still suck, then that's the worst possible combination of analytics as excuse and analytics as aesthetic killer. 

[laughedatbytime]

On 6/16/2022 at 7:28 AM, edhunter said:

I get why they're used. I recently watched Moneyball and really liked it. But it makes the game unbearable to watch. Hopefully things like pitch clocks, eliminating the shift and cutting down on pitching changes will counteract them.  They're effective over a long sample size, but in a short series, you still need to rely on baseball instincts. Tampa blew the World Series a few years ago because of over-reliance on them. 

 

 Another potential downside to them is this: My behated  Red Sox are owned by John Henry. He views the team strictly as a part of his sports business empire. He also owns the Pittsburgh Penguins and Liverpool FC. They use analytics to ensure a certain degree of success, and profit.   Last year the Red Sox made the American league championship series. Before the trading deadline they ran over 50 analytical models which showed that no combination of players on the market would have gotten them to the World Series and won, so they declined to make the investment in them. They passed on Max Scherzer, and Anthony Rizzo and settled for Kyle Schwarber.  They used analytics to choose maximum profitability over a chance for a  championship.  The revenue of 2 more home playoff games and the prestige  of a potential championship was not worth the risk of the extra investment to secure them.   It may have been their last best chance, because several of their best players are free agents this year,  and it looks like there's no way they're going to spend the money to keep all of them.  They have taken the focus away from the game and put it on the profits derived from the game. The results of the game are irrelevant  past a certain profit point.

A couple of comments:

 

Tampa was not likely to win the series (or even necessarily the game) even if they didn't pull Snell in Game Six.   Secondly, they were there, having beaten teams with far greater resources, by being smarter over a much larger sample of games.   It seems bizarre to attribute their failures to analytics while not crediting their successes to it.  Sounds like a narrative looking for an anecdote.

 

Your example of the Red Stockings is more of an indictment of the motivation of ownership than an indictment of analytics.  If a bunch of grizzled scouts would have made the same arguments while ruining the carpet by spitting tobacco juice all over it based on the eye test, and the owner considered that the best information he could get at the time and had the same mitigation the decision would have been the same, but based on inferior knowledge, except maybe regarding how much to pay to clean the carpet.