Richard keeps intimating that I have said that both Mike's system and the current system (at 3.5) would produce similar results. Yes and no. The top of the range looks similar but many of the names, and certainly the relative ranks of the goalies within that range, are very different and the lower ranked goalies are considerably less valuable under the current system. That may be a result that we think is fine but, as I have said, it does result in goalies who had statistically poor seasons getting much higher pts and relative rankings under Mike's proposed system.
Here are the top 24 Goalies
Under Current Rule with the baseline set at 3.5
Evgeni Nabokov 103.38
Martin Brodeur 102.74
Henrik Lundqvist 91.12
Roberto Luongo 79.02
J.-S. Giguere 76.13
Marty Turco 71.98
Niklas Backstrom 67.61
Manny Legace 66.60
Ryan Miller 64.13
Ilya Bryzgalov 63.85
Pascal Leclaire 62.21
Cristobal Huet 59.96
Miikka Kiprusoff 59.37
Tim Thomas 59.04
Chris Osgood 56.61
Tomas Vokoun 55.09
Martin Biron 53.67
Jose Theodore 53.49
Dominik Hasek 53.27
Cam Ward 49.13
Vesa Toskala 48.60
Dan Ellis 43.09
Rick DiPietro 42.01
N. Khabibulin 41.93
And here are the top 24 goalies
Under Mike's proposed system:
Evgeni Nabokov 106
Martin Brodeur 98
Henrik Lundqvist 94
Miikka Kiprusoff 90
Roberto Luongo 85
Ryan Miller 85
Cam Ward 83
J.-S. Giguere 80
Niklas Backstrom 78
Vesa Toskala 75
Martin Biron 74
Marty Turco 73
Tomas Vokoun 72
Cristobal Huet 71
Manny Legace 67
Martin Gerber 66
Ilya Bryzgalov 65
Tim Thomas 65
Pascal Leclaire 63
Chris Osgood 62
Jose Theodore 62
Dominik Hasek 62
Rick DiPietro 62
Mathieu Garon 57
Finally, to add some more fuel to the discussion, here is the difference between the two. Mike's proposal adds pts to every goalie - save M.Brodeur who loses pts - but the differences are telling.
Additional pts goalies get under Mike's system vs the current system with a 3.5pt baseline:
Cam Ward 33.88
Johan Hedberg 30.72
Miikka Kiprusoff 30.63
Vesa Toskala 26.40
Olaf Kolzig 25.99
Johan Holmqvist 24.46
Martin Gerber 24.44
Ryan Miller 20.87
Martin Biron 20.33
Rick DiPietro 19.99
Mathieu Garon 19.79
Chris Mason 19.08
Ray Emery 17.58
Dwayne Roloson 17.45
Tomas Vokoun 16.91
Jason LaBarbera 16.83
Carey Price 16.20
Kari Lehtonen 15.93
Patrick Lalime 14.28
Antero Niittymaki 13.00
Ty Conklin 12.21
N. Khabibulin 12.07
Dan Ellis 11.91
Cristobal Huet 11.04
Niklas Backstrom 10.39
Josh Harding 10.34
Dominik Hasek 8.73
Jose Theodore 8.51
Alex Auld 8.47
Marc-Andre Fleury 7.79
Peter Budaj 6.36
Roberto Luongo 5.98
Tim Thomas 5.96
Chris Osgood 5.39
J.-S. Giguere 3.87
Mike Smith 3.49
Henrik Lundqvist 2.88
Evgeni Nabokov 2.62
Fredrik Norrena 2.52
Ilya Bryzgalov 1.15
Marty Turco 1.02
Pascal Leclaire 0.79
Manny Legace 0.40
Martin Brodeur -4.74
The top-end goalies (Nabokov, Lundqvist, Luongo, Brodeur) don't gain much under Mike's system - they are top-end either way - but goalies like Ward, Hedberg, Kipper and Toskala are much more valuable under Mike's system. I think our current system, as imperfect as it is, does a much better job of rewarding the real-world top goalies than does Mike's system.
Subscribe to:
Post Comments (Atom)
16 comments:
Thanks for putting up the stats Doug.
For me the issue with Mike's suggestion is not the results it produces, but that it isn't directly related to goaltender performance whereas our current system is.
Even simpler systems than Mike suggests could be constructed or devised, but if they aren't actually measuring goaltender performance why use them?
Whats more the standard should be that it measures goaltender performance BETTER than our current system does, which from the evidence above, it clearly does not.
The numbers I gave from last year, and the previous 2 years were based on 4.0, which we used since the lockout.
You will not get a proper comparison of the two systems using previous years if you change from 4.0 to 3.5.
We did not use 3.5 last year. We used 4.0, so please do not try and compare what I gave with what we would have seen had we used 3.5 last year, which we did not.
The only way you are going to get an accurate representation of goalie performance is to use save percentage instead of goals against. Even then, there is always going to be the team aspect.
Goalies do not play by themselves, they are part of, and have a team in front of them. Hockey is a team based sport.
On a day to day basis, the goalie is only as good as the team infront of him.
My system increases the number of worthwhile goalies, and makes backup goalies relevant. GMs have more options now if you have a playable backup. Few, if any, draft a backup with the intention of ever dressing thier backup, except as a body on the roster in case of injury.
Hi Mike,
I'm comparing your system to the current system with a 3.5 baseline because those are the two contending options at the moment.
While I like the 3.75 baseline, I'm pretty much on my own in favour of it so it really isn't an option. Nobody, to my knowledge, has suggested we keep the current formula while maintaining the 4.0 baseline - so it really is pointless to use it as a comparison.
You are absolutely right that there are more goalies with "good" numbers but the consequence is that GMs who have high-pointing goalies under the current system will see their relative value drop significantly. The change may be a good idea or a bad idea, but it is hardly a neutral one.
Imagine if we decided to value goals at 2pts and assists at 1/2pt. The overall pts of FunHL players would not change much - but Ovechkin's GM would be much happier about the change than Crosby's GM.
Reducing the spread of good to bad goaltenders, lowers their overall value and will also reduce trading. Also by going with wins, losses and shutouts it makes this pool more like every other pool.
Gaah!! Rob!!
I think that comment single-handedly identifies why I find Mike's system so appealling--and please forgive me if I go a little off the deep-end with my zealotry in disagreeing with you--but I'm convinced that the result on the vote is going to be something obnoxious like 7-5, enough to show we have a problem: not enough to get change.
I'll deal with your second point first "makes us too much like every other pool," is, among other things, a really solid argument for having a car with square tires. It will certainly make you unique. It manages to ignore things like "efficiency" or "simplicity." And it makes hash of anyone who's in the pool for, you know, cameraderie and fellowship. The rules are only rules, and they really should serve simplicity whenever possible (though, I concede, accuracy is also important). The pool should survive no matter what the rules may be, because, like a clock, we are more than the sum of our rules.
Your first point is a logical non-sequitur, and handily explains why I turn down so many of your trade offers. Increasing the total difference in scores of a narrow pool of players (goalies being the narrowest pool--only 12 get played each week) decreases the odds of people trading the ones up at the top, and because each team has to play one each week, there just won't be much motion in the pool, since only those at the low end have an incentive to ditch.
To that end, because bad goalies carry a significant risk of netting negative points, people are putting injured goalies in their lineup--something no NHL team would do--but makes a lot of sense in our system. A 0 is better than a negative. (Doug's reporting does not include how often a number less than 1 came up, only numbers less than 0.)
By making "bad" goalies a bit more valuable, people will be more willing to trade them because it's an easy currency, and it's easier to take a risk. People look to match a net-gain with a net-loss, by making mediocre goalies a bit more valuable, it's easier to trade into or out of them.
It's why stock-splits are such an important part of the stock market: people won't move on exceptionally valuable stocks, say like Warren Buffetts Berkshire-Hathaway where each stock is worth $300 K. That's a long term investment that you don't touch. But put stocks in the $20 to $120 range and they move much faster and more frequently, because people don't perceive a big risk in changing their portfolio. (Hence why I'd bet that mid-low-end forwards and the Ds probably see more trades than any other position--the are common and all worth about the same amount)
If what you want are more trades, then what you will support is a scoring system that brings the total range of goalie value closer together, and worth the same approximate amount as the forwards that are kept at the WD. Arguing for more diverse values will decrease the total number of trade offers that will be accepted, because the risk inherent in losing a top 4 goalie becomes overwhelming, and GMs will be very conservative about engaging in such a trade.
People don't like making losing trades, they like making roughly equal trades, without a long-term disastrous risk components.
Anyway, 7-5 sound about right for the vote split on Mike's goalie proposal to anyone else?
Fisking Richard:
Richard: Gaah!! Rob!!
Cameron: Gaagh is best served live.
Richard: I think that comment (it makes the pool like many others) single-handedly identifies why I find Mike's system so appealling
Cameron: and the fact you find this appealing (our pool being just like others) is what many of us old guard find appalling.
Richard: --and please forgive me if I go a little off the deep-end with my zealotry in disagreeing with you--
Cameron: I forgive you.
Richard: but I'm convinced that the result on the vote is going to be something obnoxious like 7-5, enough to show we have a problem: not enough to get change.
Cameron: I think you overestimate the strength of opposition to the current rules. By a factor of 2.
Richard: I'll deal with your second point first "makes us too much like every other pool," is, among other things, a really solid argument for having a car with square tires. It will certainly make you unique.
Cameron: Rob's point about uniqueness shouldn't be dismissed so quickly. Prior to the FUNHL most of our gms (if not all) had been in other mundane type pools (if your office was anything like mine it will no doubt have one for the playoffs), but upon arriving in the FUNHL found its differences to be worthwhile.
Richard: It manages to ignore things like "efficiency" or "simplicity."
Cameron: Undoubtedly so, our rules are more complex than those proposed by Mike. However, unlike Mike's proposal our current system actually fairly represents goaltender value - Mike's system doesn't even pretend to do this eschewing goaltender stats (like saves, goals allowed, minutes played) entirely. It may be simpler and more efficient, but it isn't goaltending.
Richard: And it makes hash of anyone who's in the pool for, you know, cameraderie and fellowship.
Cameron: You are going have to explain this one to me. How is it that moderately complex but accurate goaltender stats interferes with camaraderie and fellowship?
Richard: The rules are only rules, and they really should serve simplicity whenever possible (though, I concede, accuracy is also important).
Cameron: Simplicity is not and never has been the hall-mark of the FUNHL. Fairness, accuracy, and fun, but never simplicity.
Richard: Your first point is a logical non-sequitur, and handily explains why I turn down so many of your trade offers.
Cameron: K - that was funny, but Rob actually does have a point. If you move to a system that merely rewards getting a start for the majority of points that a goaltender earns, you get a compression of stats that leads to bloated average class in the middle, and this bloating does in fact reduce the overall value of elite goaltenders (Rob's actual point).
Richard: Increasing the total difference in scores of a narrow pool of players (goalies being the narrowest pool--only 12 get played each week) decreases the odds of people trading the ones up at the top, and because each team has to play one each week, there just won't be much motion in the pool, since only those at the low end have an incentive to ditch.
Cameron: Funny, but without a bloated middle class to stimulate goaltender trades at the top-end, we nevertheless say the trading of: Brodeur FP, Lundqvist FP, Luongo FP, and Nabokov. In other words there isn't exactly a problem with goaltenders being locked down by a few GMs.
Richard: To that end...
Cameron: Worth repeating that the end in concern here is a straw man.
Richard:.. because bad goalies carry a significant risk of netting negative points
Cameron: A risk you yourself noted was insubstantial.
RicharD:... people are putting injured goalies in their lineup--something no NHL team would do--but makes a lot of sense in our system.
Cameron: Remind me again, which Gm played injured goaltending rather than worry about a loss of points?
Richard: A 0 is better than a negative. (Doug's reporting does not include how often a number less than 1 came up, only numbers less than 0.)
Cameron: Curiously, a number less than one but greater than zero is still greater than zero.
Richard: By making "bad" goalies a bit more valuable, people will be more willing to trade them because it's an easy currency, and it's easier to take a risk.
Cameron: Just a reminder, goaltender trades weren't a scarce commodity this year.
Richard:If what you want are more trades, then what you will support is a scoring system that brings the total range of goalie value closer together, and worth the same approximate amount as the forwards that are kept at the WD.
Cameron: Actually I suggest we merely use a statistical range that keeps starting goaltenders in line with starting first line centers. Higher than this and a goaltender requires a king's mortgage to aquire. Go to low and we have starting goaltenders who score like 2nd tier defensemen. But at 3.5 we get it just right.
Richard: People don't like making losing trades, they like making roughly equal trades, without a long-term disastrous risk components.
Cameron: Agreed.
Richard: Anyway, 7-5 sound about right for the vote split on Mike's goalie proposal to anyone else?
Cameron: I put it at 9-3 or 10-2.
I do not need simplicity, I use a computer.
Also when your goaltender is playing your favourite team, it is easier to cheer for your goaltender to loose a tight 1-0 game than to have him win it for 2 points as then your favourite team does not get the win. I hate having to bet against my favourtie team winning.
Richard,
FYI, last season there were 12 times where a goaltender got 0.00pts for a week. Four of those times the goaltender was dressing for the Lost Boys so I can see where you feel this may be a bigger problem than others do. I can not tell how many times this was the result of a GM dressing a known injured goalie with the hope of not losing pts as opposed to playing a goalie thinking he was uninjured or that he might return before the week was over.
I have found myself in the crappy position of dressing a goalie for 0 pnts because of injury on a few occasions.
Is not dressing a goalie for zero points instead of playing a risky goaltender a combination of bad drafting and an unwillingness to trade for another goaltender.
Careful there Mr. Wolf - drafting Avery and Shanahan in the first and second rnds...
I attempted to remidy things like that by getting Biron and Vokoun, lest you forgot.
I am always interested in making deals, just not crazy deals that some, ahem, like to offer.
And now the wait continues for DC to decide goalie stats AND also procedure for new rule proposals.....here's hoping we don't discuss rules at all on the ED day - takes away from the fun if there is a GM(s) upset that a rule changed / didn't change.
Why exactly do we need a new proposal for how to deal with rule changes?
Sorry - I didn't word that well. I meant figuring out a deadline for new rule proposals (or maybe I didn't look close enough and it already says this in the covenant?)
Post a Comment