All Purpose Analytics and Extended Stats Discussion

[QuadrupleDeke]

The point you're making is clear to me and I think some of us have even discussed it in this thread. I've talked about it in recent GDT when posters complain about high possession but no results. Better teams do good things that result in shots and goals. You don't get better or get goals just by shooting.

The stat is only valuable if the players are largely ignorant of it. I believe they do know about it, but their understanding is probably limited. Most will say they've heard of it but don't get into it much. Some may be lying, but I'm guessing what most players are going to come away with is "shooting is good" and they will do what they can to generate shots. Professional athletes in all sports are aware of stats to some degree, even if they don't admit it. Their careers often depend on it, but you can't play effectively when micromanaging stats. So you simplify.

In the 2 on 1 scenario you describe it may even be an unconscious decision to take the shot (even if the better play is to attempt the pass due to defender and goalie position) because of that simplified, positive association.

You're essentially describing what we could call Shrodinger's Shot. The observation of the stat is likely affecting the stat, creating a paradox. Or something.

Exactly.

I think players are more aware of it than you think, though. In the NBA, the huge push towards high shooting percentage has actually resulted in players being hesitant to take end-of-quarter buzzer beaters. They'll either not take a shot at all, or wait until just after the buzzer to release the ball. In the NFL, there have been quarterbacks who don't want to throw the ball away and would rather take a sack, since it doesn't hurt their quarterback ratings.

 

[g00n]

Exactly.

I think players are more aware of it than you think, though. In the NBA, the huge push towards high shooting percentage has actually resulted in players being hesitant to take end-of-quarter buzzer beaters. They'll either not take a shot at all, or wait until just after the buzzer to release the ball. In the NFL, there have been quarterbacks who don't want to throw the ball away and would rather take a sack, since it doesn't hurt their quarterback ratings.

On further review "Shrodinger's Stat" is probably better, but you get the idea.

I hadn't heard about the examples you mentioned but it wouldn't surprise me. When I was playing a lot of organized sports many of my coaches would try to keep the players away from the stat sheets. Hell, we had guys in HS basketball games that would go up to the scorer's table during the game to check their stats. Good way to earn the whole team some suicides the next day in practice.

 

[QuadrupleDeke]

On further review "Shrodinger's Stat" is probably better, but you get the idea.

I hadn't heard about the examples you mentioned but it wouldn't surprise me. When I was playing a lot of organized sports many of my coaches would try to keep the players away from the stat sheets. Hell, we had guys in HS basketball games that would go up to the scorer's table during the game to check their stats. Good way to earn the whole team some suicides the next day in practice.

Totally forgot about double-doubles and triple-doubles in basketball. If I were a basketball coach, I definitely wouldn't let my players know their stats in-game. Need that extra assist to complete your double double? Overpass. Need that extra rebound for your triple-double? Brick your layup and get your own rebound.

Yeah, Schrodinger's Stat is a pretty descriptive name for the Corsi phenomenon.

 

[g00n]

Totally forgot about double-doubles and triple-doubles in basketball. If I were a basketball coach, I definitely wouldn't let my players know their stats in-game. Need that extra assist to complete your double double? Overpass. Need that extra rebound for your triple-double? Brick your layup and get your own rebound.

Yeah, Schrodinger's Stat is a pretty descriptive name for the Corsi phenomenon.

The line is also pretty thin here. When we're talking about a pretty small number of shots/data per game per player based on average league numbers and the difference between good and bad is a few percentage points around a coin flip, a couple of these decisions every few games will affect the results. By the numbers, it has to.

So yeah you're right, there probably are guys who are already starting to blow the curve intentionally.

 

[ChibiPooky]

The point you're making is clear to me and I think some of us have even discussed it in this thread. I've talked about it in recent GDT when posters complain about high possession but no results. Better teams do good things that result in shots and goals. You don't get better or get goals just by shooting.

The stat is only valuable if the players are largely ignorant of it. I believe they do know about it, but their understanding is probably limited. Most will say they've heard of it but don't get into it much. Some may be lying, but I'm guessing what most players are going to come away with is "shooting is good" and they will do what they can to generate shots. Professional athletes in all sports are aware of stats to some degree, even if they don't admit it. Their careers often depend on it, but you can't play effectively when micromanaging stats. So you simplify.

In the 2 on 1 scenario you describe it may even be an unconscious decision to take the shot (even if the better play is to attempt the pass due to defender and goalie position) because of that simplified, positive association.

You're essentially describing what we could call Shrodinger's Shot. The observation of the stat is likely affecting the stat, creating a paradox. Or something.

Schrödinger's.

You said: "Better teams do things that result in shots and goals." That's true. But there's a relationship between shots and goals, and between both and winning. Shot metrics are predictive - they'll indicate that a team is likely to rebound or fall off a cliff, etc. Frustration likely comes from the expectation that they'll rebound and the reality that they haven't yet. The fact that they're predictive means that they really can't be used as negotiating tools in the traditional sense. Smart teams will look at a guy who had a "unsustainably bad" contract year and sign him to a cheap contract based on his boxcar stats, which are the only ones allowed in arbitration and therefore are really the only ones which carry weight. They might also be inclined to let someone who had an "unsustainably good" contract year walk.

Edit: could this paragraph be any more topically haphazard?

Exactly.

I think players are more aware of it than you think, though. In the NBA, the huge push towards high shooting percentage has actually resulted in players being hesitant to take end-of-quarter buzzer beaters. They'll either not take a shot at all, or wait until just after the buzzer to release the ball. In the NFL, there have been quarterbacks who don't want to throw the ball away and would rather take a sack, since it doesn't hurt their quarterback ratings.

These cases are different, though. Not taking buzzer beaters doesn't make shooting percentage a less useful stat. Same with the QBR example. In hockey, playing to take any shot instead of the best shot fundamentally changes the game.

 

[g00n]

But there's a relationship between shots and goals, and between both and winning. Shot metrics are predictive - they'll indicate that a team is likely to rebound or fall off a cliff, etc. Frustration likely comes from the expectation that they'll rebound and the reality that they haven't yet. The fact that they're predictive means that they really can't be used as negotiating tools in the traditional sense.

The whole point of this angle is that the above may not be true if the players are aware of it. Arbitration is not the only measure of success. If fancy stats = ice time and line assignments = chance at stats that get contracts/$$$$ then fancy stats matter to players.

In hockey, playing to take any shot instead of the best shot fundamentally changes the game.

Again, part of the point.

 

[g00n]

And I will add that COACHES may even potentially abuse the stat if they're not already doing it. A coach with a bad team but good "possession" stats can blame bad luck or whatever, and argue for more patience from his bosses. He can encourage his players to fire away to boost those shot totals.

 

[ChibiPooky]

The whole point of this angle is that the above may not be true if the players are aware of it. Arbitration is not the only measure of success. If fancy stats = ice time and line assignments = chance at stats that get contracts/$$$$ then fancy stats matter to players.

Actually, arbitration is the gold standard as to which stats are allowed in negotiations and which aren't. So yes, it's important. And I guarantee the players are aware of which stats are their moneymakers.

 

[ChibiPooky]

And I will add that COACHES may even potentially abuse the stat if they're not already doing it. A coach with a bad team but good "possession" stats can blame bad luck or whatever, and argue for more patience from his bosses. He can encourage his players to fire away to boost those shot totals.

If none of the players ever talk to the GM, that might work.

 

[g00n]

In fact, you might say that such a phenomenon could lead to teams focusing highly on collapsing in the defensive zone and blocking as many shots as possible, BUT trying to be aggressive in the offensive zone in order to shoot a lot and even forecheck when possible.

In other words, exactly what Trotz has been doing, which has been producing high possession stats but increasingly few wins (along with a lot of "lucky" goals against due to traffic in front, and few guys on our side who are willing to bang in front of the net to create screens, instead favoring the gliding/fancy play while looking to shoot).

Not saying this is intentional, but...

 

[ChibiPooky]

Again, part of the point.

Sure, but because the stat is no longer valid, it won't hold weight in negotiations.

 

[ChibiPooky]

In fact, you might say that such a phenomenon could lead to teams focusing highly on collapsing in the defensive zone and blocking as many shots as possible, BUT trying to be aggressive in the offensive zone in order to shoot a lot and even forecheck when possible.

In other words, exactly what Trotz has been doing, which has been producing high possession stats but increasingly few wins (along with a lot of "lucky" goals against due to traffic in front, and few guys on our side who are willing to bang in front of the net to create screens, instead favoring the gliding/fancy play while looking to shoot).

Not saying this is intentional, but...

Now you're just getting on a roll and leaving facts behind. Shot-attempt-based metrics are just that - based on attempts (not actual shots on goal). It's only Fenwick that eliminates the blocked shots.

You may also want to read Prewitt's article that I posted in some thread. The entire practice today was spent on getting guys to the front of the net (and net front battles in general).

Never mind that Trotz probably has the longest leash of any coach in the league right now.

 

[g00n]

If none of the players ever talk to the GM, that might work.

The coach wouldn't have to tell the players what his motives are. There are plenty of ways to disguise it in already existing hockey maxims about getting the puck to the net, getting shots through, making the goalie work, eventually they'll start going in, etc.

And what player is going to go to the GM and blab on the coach if his stats are high based on this philosophy? I can't imagine a player going to a GM and saying "we're shooting a lot and losing because the coach just wants to boost our stats as an excuse in case we lose...I have no proof for this and he's never said it, but that's what I feel is happening". It will be the coach's word against he players and the coach will always say "of course I want them to shoot...you can't score goals if you don't shoot".

 

[g00n]

Now you're just getting on a roll and leaving facts behind. Shot-attempt-based metrics are just that - based on attempts. It's only Fenwick that eliminates the blocked shots.

You may also want to read Prewitt's article that I posted in some thread. The entire practice today was spent on getting guys to the front of the net (and net front battles in general).

Never mind that Trotz probably has the longest leash of any coach in the league right now.

Calm down. I said it wasn't intentional. It's probably a coincidence. But the tactics and the results match, regardless of Fenwich or Corsi. It shows it can happen. You can have great possession stats, possibly through philosophy, and still not win.

 

[ChibiPooky]

The coach wouldn't have to tell the players what his motives are. There are plenty of ways to disguise it in already existing hockey maxims about getting the puck to the net, getting shots through, making the goalie work, eventually they'll start going in, etc.

And what player is going to go to the GM and blab on the coach if his stats are high based on this philosophy? I can't imagine a player going to a GM and saying "we're shooting a lot and losing because the coach just wants to boost our stats as an excuse in case we lose...I have no proof for this and he's never said it, but that's what I feel is happening". It will be the coach's word against he players and the coach will always say "of course I want them to shoot...you can't score goals if you don't shoot".

You must think players are pretty dumb

If the goals aren't going in, nobody's stats are high. Unless in the scenario you're using, Corsi is more important than goals or assists. In that case I'd question the validity of your scenario.

 

[g00n]

Sure, but because the stat is no longer valid, it won't hold weight in negotiations.

What stats are used in negotiations now? Not just arbitration, but negotiations?

Are they all 100% valid and purely the product of the player's ability and merit? How long have those stats been around? Why would the shiny new toy being used by all these teams suddenly be LESS important when all these other stats have persisted for, decades and decades?

 

[ChibiPooky]

Calm down. I said it wasn't intentional. It's probably a coincidence. But the tactics and the results match, regardless of Fenwich or Corsi. It shows it can happen. You can have great possession stats, possibly through philosophy, and still not win.

Over a small sample of games, yes. We've seen that the predictions can be off even over a full season sample (rarely). But the predictive value is much better than any other stat we have.

 

[g00n]

You must think players are pretty dumb

If the goals aren't going in, nobody's stats are high. Unless in the scenario you're using, Corsi is more important than goals or assists. In that case I'd question the validity of your scenario.

If you're losing a lot of 3-4 games and generating a lot of shots while limiting shots against you're going to have good stats. Isn't that what the Caps have been doing?

 

[Hivemind]

This discussion is just further proof of why you need a good analytics guy working with your coaching staff. The analytics guru helps the coach understand the important takeaways, who then passes the appropriate emphasis down to his players. It prevents using the stats backwards.

It would be interesting to see if the recognition of these stats has impacted player behavior in any measurable way, but I'm skeptical it has. The players who have spoken up about these stats have largely rejected them (see Doughty). Heck, even Trotz has shown his skepticism (although he often implements the ideas they support).

 

[g00n]

Over a small sample of games, yes. We've seen that the predictions can be off even over a full season sample (rarely). But the predictive value is much better than any other stat we have.

So now we're falling back on "best available" like we always do once we've looked at the reliability and determined that there's a problem.

Nobody is seriously saying fancy stats are inferior to something like +/-. That's beside the point when discussing the actual ins and outs of the stats, and what they really mean.

 

[g00n]

This discussion is just further proof of why you need a good analytics guy working with your coaching staff. The analytics guru helps the coach understand the important takeaways, who then passes the appropriate emphasis down to his players. It prevents using the stats backwards.

It would be interesting to see if the recognition of these stats has impacted player behavior in any measurable way, but I'm skeptical it has. The players who have spoken up about these stats have largely rejected them (see Doughty). Heck, even Trotz has shown his skepticism (although he often implements the ideas they support).

Which is exactly what someone who is abusing them and gaming the system would say!!

 

[ChibiPooky]

What stats are used in negotiations now? Not just arbitration, but negotiations?

Are they all 100% valid and purely the product of the player's ability and merit? How long have those stats been around? Why would the shiny new toy being used by all these teams suddenly be LESS important when all these other stats have persisted for, decades and decades?

As far as I know it's just the boxcar stats. G, A, P, TOI, +/-, etc.

A goal will always be a goal. Teams trying to score more goals doesn't change anything about that stat or about the game. Do we really have to go through this?

Never mind that Corsi/Fenwick, being predictive stats, already aren't really useful for telling you how well someone played in the past - only how likely it is that they'll continue that performance in the future.

 

[ChibiPooky]

If you're losing a lot of 3-4 games and generating a lot of shots while limiting shots against you're going to have good stats. Isn't that what the Caps have been doing?

Yes. And would you argue that they're losing because of the play of the skaters?

 

[ChibiPooky]

So now we're falling back on "best available" like we always do once we've looked at the reliability and determined that there's a problem.

Nobody is seriously saying fancy stats are inferior to something like +/-. That's beside the point when discussing the actual ins and outs of the stats, and what they really mean.

No, we're not. Again, these stats don't tell you much about the past. There is no problem here. Nobody is arguing that Corsi etc. are 100% predictive. But if you bet the Corsi line over the full season, you'll make a lot of money.

 

[Hivemind]

Re: negotiations

I believe that the NHL stopped reporting sv% for a while in the 60s or 70s because teams didn't want it to be used in contract negotiations. I think only the NHL published stats are used in arbitration.