Setoguchi benched, press boxed, rejects Contract extension

[Wildman]

Where we may have the biggest trouble, I suspect, is in purely subjective ratings like SK, EX, LD. I wonder if going forward it would be worth simply cribbing from something like EA hockey for those ratings, and using our formulas for all the quantifiable stuff?

EX should be based on number games played in the NHL.

 

[Ohio Jones]

If you rate like this, how do you account for exceptional seasons? There would be no difference between Ovechkin's 65 goal season in 07-08 and Crosby's 52 goal season in 09-10.

Good question. My initial instinct was to range the spread between 50 and 95, and then keep 96-99 in reserve for exceptional performances of the sort you describe.

In aggregate, though, it doubt it would matter all that much either way, particularly if we continue to use a three-year sliding scale in our calculations (50/25/25 or whatever). So long as the resulting ratings give us an appropriate representation of where the player ranks versus other players in the pool, I'm not convinced it particularly matters what the ratings themselves are, so long as the guy who probably should score the most actually does more often than not -- within the limitations of the sim, of course.

For example, we've discovered in this version of the sim that players with SK below certain levels appear to be far less effective in a number of respects than more fleet-footed specimens, so this would all have to be subject to testing, just as blueliners' PA and SC ratings were.

Also, Matthew's repeated simming of individual game days recently shows a very, VERY high degree of randomness in outcomes from one run to the next. As it should be, I suppose - if we could write the script for a given game in advance, there'd be no point playing it! But I think it's important to recognize that whatever we do to try and normalize or quantify the ratings process, the actual outcomes of any given season will most likely vary significantly from expectations, purely because of the random factor.

 

[Dr.Sens(e)]

There are also some great stats you can glean out of behindthenet. I am a believer in using relative Corsi as part of a formula for defensive rating (along side what we do already, such as pk minutes per game etc.). I actually reached out to the owner there to get some ideas for formulas for different ratings.

For scoring, it also has goals / 60 minutes, which could be part of the formula to augment goals / game. Same goes for primary assists versus second assists, in terms of part of the rating for passing (a guy with a higher % of primary assists gets a bump).

There are also some interesting stats like non-coincidental penalties taken per 60 minutes. That would be a great one for discipline, because the coincidental stuff shouldn't really factor in. But again, some if this has to be tested. Also depends on what sim we use too.

The key step is to do re-rates early in the off-season and then test them against our existing settings. Will give us some feel for how to Bell curve etc.

 

[Ohio Jones]

EX should be based on number games played in the NHL.

I would agree overall. But I believe it should also reflect a player's playoff experience, international experience and other considerations all of which add to a player's toolkit of coping skills for dealing with adversity. That's ultimately what EX represents - has the player dealt with a given situation before, and do they know how to deal with it.

All of these other considerations would be used to give a player a bump up or down from where they rank in pure NHL games played, so that you might well have a 24-year-old coming off his ELC who has been on a couple of deep playoff runs and represented his country a couple of times having an EX as high or higher than a 7-year vet who is golfing every April.

This is where the depth of the stats collected is important (and I'm glad to hear Nick is working with the guys at behindthenet.com, because they have a *****in' resource!), and also why the quick-and-easy routes for calculating most stats usually end up generating unsatisfying and flawed player ratings.

Nick - have you got an intern or two working on this stuff yet?

 

[Vagrant]

I find myself really concerned about stats that are subjective like SK and ST too. I wish we had a better way of getting those across. For example, because I have no horse in this race, Patrick Dwyer is one of the best skating forwards in the conference and he has like 69 rating here in that stat. I know that rerates can be done on individual stats, but those are also assuming that you can find an outlet to back you up when online scouting reports are so rare, and vague, that it makes it incredibly challenging.

 

[Ohio Jones]

I find myself really concerned about stats that are subjective like SK and ST too. I wish we had a better way of getting those across. For example, because I have no horse in this race, Patrick Dwyer is one of the best skating forwards in the conference and he has like 69 rating here in that stat. I know that rerates can be done on individual stats, but those are also assuming that you can find an outlet to back you up when online scouting reports are so rare, and vague, that it makes it incredibly challenging.

Precisely why I was suggesting drawing on a third-party resource like EA for those stats, since they have access to greater resources I evaluate them.

Part of the problem is that the sim uses blanket or "blended" ratings. SK represents a blend of top speed, acceleration and mobility (agility and turns) - possibly strength on your skates as well. ST seems to represent physical strength, but also size, itself a blend of height and weight. Reach might arguably be factored in, affecting a player's range. So t's very difficult to reliably make a quick and objective judgement, so some subjectivity will still creep in.

In other words, let's blow up the sim!

 

[Dr.Sens(e)]

I think idea of using EA or another reputed source is a gone.

The only other idea I had was to post some polls within the respective HF team message board pages. I actually think, if you could get a decent sample size, we'd get some great ratings out of the team fans. Nut that is a lot of work, and might not pan out of the thread / polls only garner a handful of ratings.

 

[Ohio Jones]

I think idea of using EA or another reputed source is a gone.

The only other idea I had was to post some polls within the respective HF team message board pages. I actually think, if you could get a decent sample size, we'd get some great ratings out of the team fans. Nut that is a lot of work, and might not pan out of the thread / polls only garner a handful of ratings.

I think your post may have been highjacked by the autocorrect fairy. Hopefully you were saying the idea of drawing on a reputable third-party source like EA was a "good one".

As far as polling the fanbase goes, what ratings would you consider this most relevant for? Something like LD? SK? Those are the sort I'm having the toughest problem quantifying...

 

[Dr.Sens(e)]

I think your post may have been highjacked by the autocorrect fairy. Hopefully you were saying the idea of drawing on a reputable third-party source like EA was a "good one".

As far as polling the fanbase goes, what ratings would you consider this most relevant for? Something like LD? SK? Those are the sort I'm having the toughest problem quantifying...

You are correct. Gone = good one, evidently.

SK, LD, CK, PH and even DF. I think PH, CK and DF should all have a significant quantifiable component to them, but could have say, 25% of the rating be subjective, because stats have a tough time telling the story for those completely. But LD and SK are two that definitely have very little data or stats to base them on. I think LD can get a bit of a boost by things like total playoff games, playoff series won, having a C or A and perhaps GW goals or better ppg in playoffs or something. But that can't be the whole story.

 

[Ohio Jones]

You are correct. Gone = good one, evidently.

SK, LD, CK, PH and even DF. I think PH, CK and DF should all have a significant quantifiable component to them, but could have say, 25% of the rating be subjective, because stats have a tough time telling the story for those completely. But LD and SK are two that definitely have very little data or stats to base them on. I think LD can get a bit of a boost by things like total playoff games, playoff series won, having a C or A and perhaps GW goals or better ppg in playoffs or something. But that can't be the whole story.

Some good suggestions there. The challenge will always be in coming up with a formula that's relatively easy to acquire data for and maintain. An interesting puzzle, to be sure.

 

[Fooladelfia]

Some good suggestions there. The challenge will always be in coming up with a formula that's relatively easy to acquire data for and maintain. An interesting puzzle, to be sure.

Matt Cooke has 59 in discipline and Doughty 51 lol........Steve Ott 50
Houston we have a problem with ratings........

 

[Dr.Sens(e)]

Some good suggestions there. The challenge will always be in coming up with a formula that's relatively easy to acquire data for and maintain. An interesting puzzle, to be sure.

Yeah, the thing is the subjective ones really only need to be overhauled once a year. Things like LD, SK and a few others don't really change that much year to year. They are slower moving / trending. Of course, if some major instance of clear LACK of leadership occured, a subjective bump down (or up I suppose) could happen on a case by case basis.

 

[Vagrant]

That is another tough thing about the ratings for me is that having DF and shorthanded minutes so closely linked can be a misnomer. There are teams rife with defensive forwards that simply can't find time for strong defensive players to crack the group or, the more important, there are certain players that teams don't like to use on the kill due to moderation of minutes.

 

[Ohio Jones]

Matt Cooke has 59 in discipline and Doughty 51 lol........Steve Ott 50
Houston we have a problem with ratings........

What a happy coincidence! We're having a discussion here about that very issue. Perhaps you'd like to chime in with your thoughts/suggestions?

Yeah, the thing is the subjective ones really only need to be overhauled once a year. Things like LD, SK and a few others don't really change that much year to year. They are slower moving / trending. Of course, if some major instance of clear LACK of leadership occured, a subjective bump down (or up I suppose) could happen on a case by case basis.

Agreed.

Since we'll still be using STHL for next season (but hopefully not beyond!), I think our easiest - and most reliable - course, especially with those more static characteristics, is to simply go 3rd party. Between EA and STHL there is a direct correlation for some ratings, others would need to be bundled. I'll see if I can suggest some formulas to combine them... then we'd need to figure out whether that gives us an absolute rating, or a ranking within a predetermined bell curve of ratings. Either way, it's going to be better than what we currently use.

That is another tough thing about the ratings for me is that having DF and shorthanded minutes so closely linked can be a misnomer. There are teams rife with defensive forwards that simply can't find time for strong defensive players to crack the group or, the more important, there are certain players that teams don't like to use on the kill due to moderation of minutes.

Agreed, nobody is especially happy with the current formula. Nick's been having discussions with the behindthenet folks about exactly those concerns. We'll see if we can't use a formula that gives weight to some other less fickle considerations (relative Corsi, etc.).

All of this doesn't touch on perhaps the least well understood part of the ratings process - the goalies. On the one hand, we shouldn't be seeing quite so many high save percentages this season; on the other hand, teams without truly elite goalies are struggling more than should reasonably be expected this season. And we also have to address the inhuman endurance and durability our goalies have shown the past two seasons. Lots to dig into in net.

 

[MatthewFlames]

Matt Cooke has 59 in discipline and Doughty 51 lol........Steve Ott 50
Houston we have a problem with ratings........

Actually those ratings are fine. Players with low DI get lots of penalties. So Ott, who is always a PIM leader should have that low rating.

 

[Ohio Jones]

Actually those ratings are fine. Players with low DI get lots of penalties. So Ott, who is always a PIM leader should have that low rating.

Doughty at 51 though? Seems a bit harsh.

The larger issue is whether we're generating an appropriate number of penalties, and what kinds they are. The DI ties into FG in a weird way that is pretty flawed. Something else to test.

 

[Fooladelfia]

Doughty at 51 though? Seems a bit harsh.

The larger issue is whether we're generating an appropriate number of penalties, and what kinds they are. The DI ties into FG in a weird way that is pretty flawed. Something else to test.

Kershaw don't know ......

 

[Dryden]

Doughty at 51 though? Seems a bit harsh.

The larger issue is whether we're generating an appropriate number of penalties, and what kinds they are. The DI ties into FG in a weird way that is pretty flawed. Something else to test.

Isn't that why we have challenges?

 

[MatthewFlames]

Doughty at 51 though? Seems a bit harsh.

*cough* ratings *cough* challenges *cough*

 

[Fooladelfia]

Isn't that why we have challenges?

I know...But how at first place he was at 51 ?? LOL

 

[MatthewFlames]

I know...But how at first place he was at 51 ?? LOL

I did it. Now go back to your vineyard...

 

[Ohio Jones]

Life at Alvaro's winery:



(Okay, I know, Spanish not Sardinian, it's all Greek to me!)

And yes, that's what challenges were for. Not everyone did a rating by rating review of their guys (their fault), and in some cases no doubt there were more screwy ratings than available challenges. Pretty sure most people didn't even think to look at things like EN and DI, where hidden problems could easily lurk - like this. With luck it won't be an issue next season.

(LOOOOVE that chase music, by the way. What an awesome movie.)

 

[Toronto_AGM_Adil]

You are correct. Gone = good one, evidently.

SK, LD, CK, PH and even DF. I think PH, CK and DF should all have a significant quantifiable component to them, but could have say, 25% of the rating be subjective, because stats have a tough time telling the story for those completely. But LD and SK are two that definitely have very little data or stats to base them on. I think LD can get a bit of a boost by things like total playoff games, playoff series won, having a C or A and perhaps GW goals or better ppg in playoffs or something. But that can't be the whole story.

I agree the concept of SK sounds impossible to quantify, however it has a real impact to certain stats in the sim. SK is about carrying the puck, higher SK results in two things:

1. Players will carry the puck into the offensive zone more often then dumping.

2. Players are harder to move off the puck while carrying.

In regards to NHL stats this has some rough correlation to giveaways (and yes I know this isn't the greatest state).I think gveaways offset by time on ice will give us a good indication of SK. Another good state if we could find it is hits taken/received. Btw, size is correlated to ST and is already factored into the sim.

Its very important to note that the SIM treats the stats with a non-linear relationship. The difference between 65 and 70 is significantly less then that of 70 to 75. Therefore its important how we distribute the average and sigma of the bell curve between ratings. To me this is the biggest flaw in our ratings today.

Btw, I personally think that no ratings should ever be subjectively determined by anyone within the league, I'd rather have flawed third party ratings and avoid any potential controversies...

 

[Ohio Jones]

Interesting comments, Adil. I think it's dangerous to read too much into giveaways, however. The people who have the most giveaways are invariably the people who "lug the mail", i.e. the people who handle the puck the longest/most. I know you suggest pro-rating giveaways for time on ice (a giveaways per 60 minutes stat), but that will still mean that PP QBs will be considered poorer skaters than snipers and power forwards who touch the puck less, and they'll all be considered poorer skaters than checkers/stay-at-homers are. I'd say the two have - or should have - little tO do wit one another.

 

[Dr.Sens(e)]

Btw, I personally think that no ratings should ever be subjectively determined by anyone within the league, I'd rather have flawed third party ratings and avoid any potential controversies...

I agree. It is my hope we will come up with sources for ratings and a formula, that has us compile the ratings, but not have to review, audit or change them individually at all, other than adjust the scale of all of them, or a bell curce kind of thing.