DF

[Chris-Wild]

Yes hit are valued different from arena to arena too (looking at NYR/MIN where you touch a guy and it is a hit), but don't think IT is nearly as important as DF.

I agree DF is important and hard to nail down but nothing is impossible and trying to find the missing link isn't going to hurt anyone.

Besides, how do u know DF is more important than any other ratings in the sim, we don't.

[anthgrt-la]

Yes hit are valued different from arena to arena too (looking at NYR/MIN where you touch a guy and it is a hit), but don't think IT is nearly as important as DF.

I agree DF is important and hard to nail down but nothing is impossible and trying to find the missing link isn't going to hurt anyone.

Besides, how do u know DF is more important than any other ratings in the sim, we don't.

Similar to how ST is more important. Bumps the various stats us by 5, 10, 15, etc and watch how they affect OV ratings. There is a reason ST affects OV as much as it does.

[jonquixote]

I think it's better if DF ratings stay pretty consistent because of how subjective it is. We've discussed this before, in the early days of the league. One reason is that unless there's a hard and fast and accessible formula, GMs can be blind-sided. "What do you mean the ratings committee dropped Eriksson's DF 10 points just because his blocked shots went down after he moved to the Bruins? WTF? The fact that the Bruins have a deeper platoon of PKers than Dallas doesn't mean Eriksson is any less effective a shutdown winger."

We can experiment with stats and formulas, but there aren't really good numbers for these things. The advanced stats conversation rages on at the highest levels of the game, with teams spending hundreds of thousands of dollars trying to work out ways to read things like this. Our stats guys can try, but it's unlikely they're going to come up with magic math to work out something that's reasonable and makes everybody even moderately happy.

It's probably a good idea to ensure that we comb over DF (and SP and SK) ratings thoroughly, especially to keep bringing them down where justified. Maybe a thread to keep track of articles we've read and incidents we observe vis a vis defensive play. Maybe a standard reduction in SP and SK players over a certain age too (subject to challenge of course).

[Nick - Philly]

I think it's better if DF ratings stay pretty consistent because of how subjective it is. We've discussed this before, in the early days of the league. One reason is that unless there's a hard and fast and accessible formula, GMs can be blind-sided. "What do you mean the ratings committee dropped Eriksson's DF 10 points just because his blocked shots went down after he moved to the Bruins? WTF? The fact that the Bruins have a deeper platoon of PKers than Dallas doesn't mean Eriksson is any less effective a shutdown winger."

We can experiment with stats and formulas, but there aren't really good numbers for these things. The advanced stats conversation rages on at the highest levels of the game, with teams spending hundreds of thousands of dollars trying to work out ways to read things like this. Our stats guys can try, but it's unlikely they're going to come up with magic math to work out something that's reasonable and makes everybody even moderately happy.

It's probably a good idea to ensure that we comb over DF (and SP and SK) ratings thoroughly, especially to keep bringing them down where justified. Maybe a thread to keep track of articles we've read and incidents we observe vis a vis defensive play. Maybe a standard reduction in SP and SK players over a certain age too (subject to challenge of course).

Well said.

[Jonathan]

I tend to use give aways/take aways in my challenges, as well as BkS, TOI. Are they bullet proof? Nope. No statistic generally is when it comes to defensive play. The problem with that is, would the rating committees accept my challenge on Pietrangelo if I said "Well the statistics may in fact be bogus, so instead, I am just going to say Pietrangelo deserves 75 DF, because, quite frankly, I think he deserves it." Is that really much of an argument? So as much as you can poke holes thru the statistics used in challenges, they are nevertheless necessary in my opinion for sim leagues, as they do show at least SOMETHING, for a stat that is already very subjective.

[Nick - Philly]

I tend to use give aways/take aways in my challenges, as well as BkS, TOI. Are they bullet proof? Nope. No statistic generally is when it comes to defensive play. The problem with that is, would the rating committees accept my challenge on Pietrangelo if I said "Well the statistics may in fact be bogus, so instead, I am just going to say Pietrangelo deserves 75 DF, because, quite frankly, I think he deserves it." Is that really much of an argument? So as much as you can poke holes thru the statistics used in challenges, they are nevertheless necessary in my opinion for sim leagues, as they do show at least SOMETHING, for a stat that is already very subjective.

Quotes are really useful in DF challenges too.

[Chris-Wild]

I tend to use give aways/take aways in my challenges, as well as BkS, TOI. Are they bullet proof? Nope. No statistic generally is when it comes to defensive play. The problem with that is, would the rating committees accept my challenge on Pietrangelo if I said "Well the statistics may in fact be bogus, so instead, I am just going to say Pietrangelo deserves 75 DF, because, quite frankly, I think he deserves it." Is that really much of an argument? So as much as you can poke holes thru the statistics used in challenges, they are nevertheless necessary in my opinion for sim leagues, as they do show at least SOMETHING, for a stat that is already very subjective.

Quotes are really useful in DF challenges too.

Agree, depending on the source i find them more useful that a bunch of numbers...

[Chris-Wild]

I think it's better if DF ratings stay pretty consistent because of how subjective it is. We've discussed this before, in the early days of the league. One reason is that unless there's a hard and fast and accessible formula, GMs can be blind-sided. "What do you mean the ratings committee dropped Eriksson's DF 10 points just because his blocked shots went down after he moved to the Bruins? WTF? The fact that the Bruins have a deeper platoon of PKers than Dallas doesn't mean Eriksson is any less effective a shutdown winger."

We can experiment with stats and formulas, but there aren't really good numbers for these things. The advanced stats conversation rages on at the highest levels of the game, with teams spending hundreds of thousands of dollars trying to work out ways to read things like this. Our stats guys can try, but it's unlikely they're going to come up with magic math to work out something that's reasonable and makes everybody even moderately happy.

It's probably a good idea to ensure that we comb over DF (and SP and SK) ratings thoroughly, especially to keep bringing them down where justified. Maybe a thread to keep track of articles we've read and incidents we observe vis a vis defensive play. Maybe a standard reduction in SP and SK players over a certain age too (subject to challenge of course).

Great perspective as always, i guess my wonder has always been 'just how important is this pesky df rating in the sim'...that's the crux of my quest for sim perfection...a non-subjective super machine that will accept raw numbers and spit out a result, a result that we can see, a result that isn't based on the 'belief' i think Franson should be a 40 df when stacked vs. Weber @ 78...maybe ill try running some test sims this summer

[Stefan - SabresGM]

I still think we need to get the original calculations down. You only have two challenges at the end of the day. But what I don't get is if a player has an off season as far as DF, he isn't really punished in the same way as a player who has an off season with SC. I know it's harder to calculate, but players should be punished for an off season, that's the idea of a NHL re-rate league. I shouldn't be able to come in here and throw some quotes down with a GM or coach saying how awesome player X is at defense.

The main point I was trying to make with the calculations is that I think we need to start accounting for the teams performance as well as the players. You obviously weight it so that the individuals performance makes up more of the rating, but you can't ignore the team as a whole.

Let's compare two players.

Zdeno Chara and Mike Weaver. Now Weaver is a decent Dman, but he definitely shouldn't have a DF rating anywhere near Chara's.

Chara:
3:13 SHTOI/G
89BkS
48GvA
18TkA

Weaver (using just his FLA stats):
2:53 SHTOI/G
104BkS
17GvA
10TkA

Now just off of those stats alone, you are pretty much saying they are nearly as good as each other. Weaver has higher BkS, Chara higher TkA. Weaver with lower GvA and Chara with more PK time per game. So if you average it out they would probably come out similar. Now a challenge might make Chara's a lot higher, but why should someone have to challenge it?

If you then add in the teams performance you get a totally different story.

Florida are 27th in the league with 3.14GA/G compared to Boston who are 2nd with 2.19. Florida has had 49 PP goals against them with Weaver leading the team on PK time. Boston has had 36 against with Chara leading the team in time. Still doesn't explain why Weaver has more Bks. Well, Florida has had 2005 shots on target against them, with Boston facing 1846. It may not seem like much, but if the trend stays the same for shots that miss/get blocked, that's an 8% difference. The difference between Weaver and Chara's BkS... 14%. So potentially 8% of those can be attributed to having more opportunity to block shots, leaving just 6% behind for random variation. Now look at this one... Chara has been on the ice for 61GA, that's more than anyone else on his team. Well obviously, he spends the most time on the ice. Take his GA vs Ice Time and for every 25 minutes he plays, the other team scores one against him. Weaver has 65GA when on the ice, which is nearly the same as Chara's, again supporting him being equal, but throw in his TOI and it becomes a goal for every 16 minutes he plays. That means if he is playing the same ice time as Chara, he would have 101GA when he was on the ice.

I'm not saying this would all be perfect, there will always be someone like Kuba who slips through, but if we can pick out the relevant stats and weight them correctly with more important ones making up a higher percentage of a players DF, we should come out with the right results. The percentages and numbers would just need a lot of work and tweaking to get to where you want the stats to sit.

[Chris-Wild]

I wonder if giveaways/takeaways is a stat only useful for forwards in DF...don't dmen have more opportunity to give the puck away since they distribute the puck and forwards receive it?

[Ben-Sens]

I wonder if giveaways/takeaways is a stat only useful for forwards in DF...don't dmen have more opportunity to give the puck away since they distribute the puck and forwards receive it?

I think Giveaways and takeaways are puck control related not defense related. Just personal opinion

[jonquixote]

I wonder if giveaways/takeaways is a stat only useful for forwards in DF...don't dmen have more opportunity to give the puck away since they distribute the puck and forwards receive it?

I think Giveaways and takeaways are puck control related not defense related. Just personal opinion

Yeah. Certain types of puck carriers almost always have a lot of "giveaways" (which are also subjectively measured, different from rink to rink) because they carry the puck more, get hit more, etc. Getzlaf is a guy who has a lot of giveaways. So are Taylor Hall and Phil Kessel. But I think it means something different when a guy like Joe Thornton has a lot of giveaways versus a guy like Kadri or Eberle. But even still, if you look at the top giveaway players they're all top line guys who create a lot of offense. It's not a stat that can be counted on to give an accurate viewing of defensive ability. Semin and Staal have the same giveaways right now, but I'm pretty sure one of those guys is okay taking an own-zone faceoff with 30 seconds left, and one is not.

[Stefan - SabresGM]

I've never been a fan of giveaways or takeaways myself to be honest. I just used it as an example. What stats we consider relavent would need to be discussed.

[anthgrt-la]

I know in a few other leagues GvA/TkA weren't even used nor allowed to be argued for DF. Purely used for PC, which I find more realistic.

[Chris-Wild]

Been doing a little research...i understand how difficult DF has been to pin down in the past...that said there are a TON of sites on the internet that complile stat upon stat...why kill ourselves trying...we could probably use a couple sources, figure out what makes sense and then develop a formula to spit out a solid DF rating.

Couple sites that do the heavy lifting:
http://www.extraskater.com/
http://www.behindthenet.ca

Then the RC could examine DF to correct any annomilies, perhaps DF is only created on full season rerates and that carries into midratings seasons since it would be more work than other ratings...isn't that a good idea?