Importance of Takeaways/Giveaways

[Like Eich]

As a guy who likes statistics, especially the obscure ones I've always been intrigued with NHL takeaways and giveaways. I know its a tough stat to track but it has to have some merrit especially with guys like Mark Stone who perennially win the steal title.

I'm also intrigued with turnover plus/minus. Not exactly standard plus/minus, since it measures how many more taleaways a player has than the amount of times they've turned it over.
The King of this statline is noneother than the magic man himself, Pavel Datsyuk. 740 takeovers, 373 giveaways for his career. Another example of a good statline for this is a guy like Toews who has a career 525 takeaways and a 210 giveaways. Bergeron is also good at 373 takeaways and 194 giveaways.

There's also the opposite of this. A guy like Tyler Seguin has 196 takeaways, and 241 giveaways. Another is a guy like Getzlaf at 371 takeaways and 596 giveaways. Joe Thornton is at 621 takeaways and 762 giveaways.

Then there's what I call the wild card category. This is for guys who can gank the puck, but also turn it over quite a bit. A good example is a guy like Tavares who is at 455 takeaways, and 459 giveaways. Another one is Steven Stamkos at 292 takeaways and 299 giveaways. Jumbo Joe could probably be placed here too. All of this is Very interesting to me...

So my question is, how important are these stats? I'd like to think they have some merit, regardless if the stat is tough to track since it's become an art for some players. Esp if their takeaways/giveaway ratio % is good. It does impact the game in a major way, and goes a long way when measuring a good player imo. What do you guys think?

 

[BlueDream]

I think they're similar to stats like blocked shots and even +/-, meaning they can be used along with other material, but it doesn't necessarily tell the whole story by itself.

You can be a good defenseman and block a lot of shots, certainly. But you can also be bad and block a ton of shots just because all you know how to do is stand in front of the puck and you're poor at actually getting it out of your zone.

Giveaways are like that. A player might be bad if they give it up a lot, but in the examples above, elite players like that will a lot of times have a lot of giveaways simply because they have the puck a lot and it's bound to happen.

Overall, I don't think they show that one player is better than another generally, unless you're in a league of your own when it comes to pick-pocketing like Datsyuk.

 

[NHLPaul]

They have some merit but giveaways in particular are a flawed stat. Leaders every year are some of the best players in the league bc they have the puck more than anyone and are always trying to create

 

[duul]

Giveaways is a terrible stat to judge players by. The difference between Mark Fayne and P.K. Subban is that Fayne is simply going to ring the puck around the wall to the winger, and not be given a turnover.

P.K. Subban is going to try and make one move and then throw a saucer pass up the middle of the ice to a breaking winger in hopes to spring him on a breakaway. All star defencemen are always looking to make these more difficult passes because they pass off far more often than a ring around the boards.

P.K. Subban leads the league in turnovers but I bet he is also one of if not the best first pass D in the NHL.

 

[G0AT]

D-men will naturally have higher turnovers, but it doesn't necessarily mean they're turnover machines or bad players.

I do think the stat has merit with Forwards though, especially if you combine it with the players possession stats. If a player has the puck a lot with low turnovers, while stealing the puck at a high rate it at least means they're not costing their team much, and creating chances at the same time. A nice thing when digging deeper into comparing players statistically.

(Btw, I think both stats are poorly recorded...)

 

[white galaxy*]

very interesting and important stat as puck possesion is obviously very important

but its a complicated stat to keep , players have to be categorized by position and strength of their team etc

but if im an nhl team im paying somebody to keep it on all players ,
ecspecially turn overs by dmen as those are the costly ones

 

[duul]

very interesting and important stat as puck possesion is obviously very important

but its a complicated stat to keep , players have to be categorized by position and strength of their team etc

but if im an nhl team im paying somebody to keep it on all players ,
ecspecially turn overs by dmen as those are the costly ones

If you look at the list of most turnovers in the NHL, dmen make up a large portion.

 

[white galaxy*]

If you look at the list of most turnovers in the NHL, dmen make up a large portion.

my point is amongst dmen whom have the least vs the most

as dmen whom handle puck well are very very valuable no matter if they dont do anything else extremely well because controling puck in your end is imo the most important factor of making a good team . without it theres hell to pay

 

[Noldo]

Turnovers are difficult stat because we do not have the context for them. If we had number for pass attempts, we could give turnovers the necessary context (if majority or large portion of a player's passes result in turnover, he may be trying to be too fancy and becomes detriment for the team).

 

[Phenomenon13]

Giveaway/takeaway per 60 would be more accurate surprised it isnt done. It would account for differences in ice time and they need to make a statistic to combine it with corsi (possession)

 

[Pavels Dog]

Both are flawed stats and especially for D-men it seems to not be useful at all. I mean Drew Doughty has one of the worst takeaway/giveaway ratios this year and he's a Norris candidate.

 

[DanZ]

Both are flawed stats and especially for D-men it seems to not be useful at all. I mean Drew Doughty has one of the worst takeaway/giveaway ratios this year and he's a Norris candidate.

They aren't stats. Maybe there is a little subjectivity to what a takeaway and giveaway is, but that doesn't make them flawed. The only thing that's flawed is the conclusions that some people draw from them. That doesn't make the stat in and of itself flawed.

 

[JWEKD]

Not only are these stats skewed for players that handle the puck more, but they also get skewed because of where the game is being played. They are very subjective stats that are interpreted differently and recorded differently by each bald, mustachio'd official NHL scorer in each arena.

 

[Trap Jesus]

I think in theory they're sort of important, although a lot can depend on how those players actually play, and the systems they're in. Just on my team (Bruins), I know that Marchand carries the puck A LOT, whereas a player like Eriksson, while technically more of a responsible player, isn't really known for holding onto it very long. Of course Marchand is going to have far more giveaways. Doesn't mean what he's doing is less effective though. I also know that if you are a poor possession player, your odds of getting takeaways increase and your odds of giving the puck away decrease simply because you don't have the puck the majority of the time.

In practice, the stat is close to useless. I don't want to say completely useless, but the fact that they're tracked extremely differently in different arenas, and the thought that the lines of what constitutes a giveaway vs. a takeaway are very blurred, it's just not a reliable stat to look at.

 

[struckbyaparkedcar]

They're decent to illuminate why another stat is good or bad, like Pavel Datsyuk having an absurd corsi% and tons of takeaways, or Tyler Ennis not correlating with ES offense and having tons of giveaways.

But mostly they're just indicators of who's most involved in play.

 

[Lonewolfe2015]

Not only are these stats skewed for players that handle the puck more, but they also get skewed because of where the game is being played. They are very subjective stats that are interpreted differently and recorded differently by each bald, mustachio'd official NHL scorer in each arena.

I wasn't sure just how skewed they were until O'Reilly changed teams.

Led the leagues for years in takeaways, a takeaway master in Colorado (objectively and statistically). Goes to Buffalo and can barely crack the top 40 in the league. Colorado is high or Buffalo is low or both.

Also brings into question Mark Stone's near 2x lead on the 2nd place Skinner.

 

[DanZ]

I wasn't sure just how skewed they were until O'Reilly changed teams.

Led the leagues for years in takeaways, a takeaway master in Colorado (objectively and statistically). Goes to Buffalo and can barely crack the top 40 in the league. Colorado is high or Buffalo is low or both.

Also brings into question Mark Stone's near 2x lead on the 2nd place Skinner.

Looking at a players takeaways in away games is probably a more reliable indicator of who is truly the best at takeaways. The 29 different arenas assuredly normalizes any sort of biases.

 

[Protest]

Looking at a players takeaways in away games is probably a more reliable indicator of who is truly the best at takeaways. The 29 different arenas assuredly normalizes any sort of biases.

Yes you would think. Like looking at ERA in away stadiums for a pitcher or power stats for a batter, but it's actually more similar to looking at the amount of errors a fielder makes in away games. That's because the stat isn't actually a stat, it's a count of a series of judgments made by someone in the arena that night. It is not definitive like a goal or an assist. So while looking at only away games is an improvement, it's still not a very good statistic, and one they should probably just drop.

 

[JWEKD]

Looking at a players takeaways in away games is probably a more reliable indicator of who is truly the best at takeaways. The 29 different arenas assuredly normalizes any sort of biases.

There is no bias for home teams. The official stat is recorded by a neutral NHL official. The stat recorded (giveaways or takeaways), in a particular building, if high or low, is that way for both teams.

Look at any game in Montreal for instance. That guy sometimes records a giveaway if the opposing team controls a puck chipped up the boards and out of the D zone. While other guys in other arenas won't record a giveaway if I literally passed the puck directly to the opposition.

I have experience with tracking stats for NHL games. Giveaways and takeaways are slightly more useless than +-. At least +- is recorded in a definitive manner.

 

[Hardyvan123]

I think they're similar to stats like blocked shots and even +/-, meaning they can be used along with other material, but it doesn't necessarily tell the whole story by itself.

You can be a good defenseman and block a lot of shots, certainly. But you can also be bad and block a ton of shots just because all you know how to do is stand in front of the puck and you're poor at actually getting it out of your zone.

Giveaways are like that. A player might be bad if they give it up a lot, but in the examples above, elite players like that will a lot of times have a lot of giveaways simply because they have the puck a lot and it's bound to happen.

Overall, I don't think they show that one player is better than another generally, unless you're in a league of your own when it comes to pick-pocketing like Datsyuk.

Pretty much this.

The obsession with stats is getting overboard IMO as much of hockey in positioning goes unnoticed by stats.

 

[DanZ]

Yes you would think. Like looking at ERA in away stadiums for a pitcher or power stats for a batter, but it's actually more similar to looking at the amount of errors a fielder makes in away games. That's because the stat isn't actually a stat, it's a count of a series of judgments made by someone in the arena that night. It is not definitive like a goal or an assist. So while looking at only away games is an improvement, it's still not a very good statistic, and one they should probably just drop.

I think it's a good statistic for what it's intended. I think that generally those players that are good at taking the puck away will compile a lot of takeaways, regardless of the stat counters. It's not any coincidence that Stone or Datsyuk have a lot of takeaways, so in the sense that it allows us to distinguish those good at taking the puck away from those that aren't so good, it does work. If two players are close in takeaways than I agree that it's not really valid to say one is better than the other, but if there is a significant gap between two players I think it's a reach to say that it's due to bias or subjectivity.

As with a lot of statistics, there is a margin of error.

 

[Serenade]

I think giveaways per touches are more important. I have a feeling the leaders in giveaways are also much lower in this stat.

Cross tabulate Giveaways/Possession vs Time on Ice. That would be interesting to look at.

 

[Pavels Dog]

They aren't stats. Maybe there is a little subjectivity to what a takeaway and giveaway is, but that doesn't make them flawed. The only thing that's flawed is the conclusions that some people draw from them. That doesn't make the stat in and of itself flawed.

Takeaways is a decent stat when used in context, puckhounds like Datsyuk, Toews, Bergeron etc. usually rack up TkA's.
Giveaways is completely useless though. Completely. Like you know how people say +/- is useless? At least good players usually end up in the + more often than not. For giveaways, the best players will look the worst, all the time.

Leaders in giveaways by year:

15-16: Subban, Doughty, Burns
14-15: Subban, Tavares, Burns
13-14: Karlsson, Hjalmarsson, Hall
12-13: Phaneuf, Markov, Byfuglien
11-12: Kovalchuk, Thornton, Getzlaf

Norris trophy winners/candidates, elite defensive d-men, elite forwards...