League Setups

[JohnnyFantasy]

Does anyone else feel like the way stat categories are chosen for fantasy hockey often doesn't make a ton of sense? Like including a bunch of peripheral categories, but not Points (so that actual scoring gets more weight)? Or including PIMs (as penalties are detrimental to a team)? I see so many people talking about really deep dynasty leagues (with salary caps, farm systems, etc.) and yet they're based on such wacky setups. Am I missing something? Is it just inertia because these have always been the categories?

[Pengwin7]

I think you'd have to be more specific about what is irking you.

Some general thoughts (based on where you seem to be going):
1. A lot of leagues now have "weighted points". This is what fantasy football has done for years. 6pts for a TD, 1pt for each 10 rushing/receiving yards, 1pt for 25 passing yards. It's a good set-up and eventually fantasy football found it's marks.
Fantasy hockey is a bit like this too. Personally, I suggest setting G=1pt, A=1pt... because that's what we've been used to for the history of hockey... then take everything else and downweight it to some decimal amount. I think a good league realizes that PIM are not really "positive", but they are fun to count. My own dynasty started this year, WHL, counts PIM and that same rate as SOG, HIT, BLK... all worth +0.05pts. A player would need 20 of any of these stats to equal a 1pt goal or 1pt assist.

2. A lot of ROTO leagues are still by "category". This can be hard because PIM is a category the same way G is a category. It overvalues PIM... so a good commissioner has to find ways to re-balance and up-weight scoring. My favourite ROTO leagues these days are counting all of G, A, and P. Some people quickly call P as "redundant" ("we're already counting it" - some say), but it is basically just upweighting Goals and Assists as a 1.5x multiplier category. I notice you have that in your SIG - so I'd say your league has a nice set-up.

3. Also noted in your SIG: SHP. This, IMO, is the absolute WORST category in fantasy hockey. Here's statistical example why:
JimmyP.Killer: On for 100 short-handed kills. Successfully kills 80, gives up 20 goals against, does register 2g,1a while successfully killing the 80 penalties.
JohnnyP.Killer: On for 100 short-handed kills. Successfully kills 90, gives up 10 goals against, doesn't score any SHP.
Q: Who's better penalty killer?
A: JohnnyP is better. His team is -10 during 100 kills. JimmyP's team is -17... but you are rewarding his 3 SHP... making him better than JohnnyP as "fantasy hockey value" goes with respect to their SH-roles.
[Besides, a league counting +/- is giving JimmyP his extra +3 at SH time, while Johnny P, the better killer, gets nothing.]
Meh. This is why SHP is a poor stat in fantasy hockey. I will argue it furiously whenever tabled in any of my leagues.
You are missing capturing player value correctly - which is sort of the goal of fantasy hockey - so don't include it. My 2 cents.

Anyways, welcome to the forums!!!
New thoughts and hockey comments are always welcome!

[JohnnyFantasy]

Thanks Pengwin.

1. Yeah, I'm in a weighted points league as well. I do find that works a lot better, although a lot of the weighting systems I have seen also seem to be off. I like your system, where the peripherals are marginalized. I think the Yahoo default is something like 0.5 per PIM, and 3 per goal, meaning someone who gets 3 minor penalties has the same fantasy value as someone who scores a goal (all other things being equal), which is crazy. Personally, I think goals should be weighted a bit more than assists (as they are rarer), but your setup sounds a lot more sensible than some others I've played in.

2. I couldn't agree with you more. In stat categories leagues, I of course want to value the peripherals, but I don't know why a lot of people don't try to balance those by adding Points. That leads to silly things like someone winning G and A, but losing SOG, HITS and BLKs and (supposing goalie and other cats were split) losing the match.

3. Point taken. The only rebuttal I can think of is that SHP isn't there to measure penalty killing, it's there to measure a special (oftentimes extraordinary) effort to score when your team is down. But, I've been falling out of favour with SHP generally since shorties occur so seldom (and thus, in categories leagues, teams will often win that category 1-0 by the random luck that they have one that week). I may be suggesting a vote in my league to remove this cat next year. What do you think about PPP? That's more of my hot-button issue. I understand it's important for teams to take advantage of their powerplay opportunities, but I have no idea why a player should get extra fantasy points for scoring when they are on the man-advantage (and thus it is easier for them to do so). The only justification I can think of is that it is another scoring-based category, and so it could be useful to include to balance out a lot of peripherals. In general though (and in points-based leagues especially) I can't see the justification. (Of course, this seems to be a default category in nearly every league ...leading to my further irking).

Thanks for the welcome. I've been looking for a good discussion regarding league setups, as I've been putting a lot more thought into mine this year, yet it seems to be something a lot of people take for granted. (Maybe I'm just not looking in the right places). I've starting blogging about it (website in my signature) to collect my thoughts, if you're interested in more ramblings, but I definitely appreciate a good back and forth debate with people on these points, so thanks for responding.

[keytolife]

In general i don´t like roto leagues that much, i prefer league with point adjusted to the individual categories wheter it be head to head or not. I also hate when stats are counted that do not really mimic value in real NHL like +/- or PIMs. Only problem with this setup is usually there is no way to value FOW fairly so i prefer it left out as well.

Thats the setup of my favorite 24 team league:

Skaters:
Goals: 18
Assists:18
Blocked Shots: 2
Games Played: 1
Game Winning Goal: 10
Hit: 1
Hattrick: 5
OTG: 15
PIM: 2
Powerplaypoints:10
SOG: 1
Shorthandedpoints: 10

Goalies:
GP:1
Saves:2
GA: -7
Overtime/Shootout Loss:5
PIM:2
Points: 20
Win: 15
Shutout: 30


For catergory leagues i like the added points system a lot and i despise +/- and PIM as stats.
This allowes D men to carry nearly as much value as forwards and multicat guys are very valueable in this setup while high ocatane offensive guys are still the most valueable.

[Pengwin7]

Only problem with this setup is usually there is no way to value FOW fairly so i prefer it left out as well.

Skaters:
Goals: 18
Assists:18...

For category leagues i like the added points system a lot and i despise +/- and PIM as stats.

Good thoughts.

From being on these boards, it was a year or two ago that I read somebody's interesting set-up regarding FOW.
Somebody had said that Fantrax allows some stats to be counted only from certain positions - like FOW can be counted only from Centers.
I thought this was a great idea, as a fantasy GM (i.e. "coach"), somebody should be allowed to get some points from FOW from a "guy taking a face-off", which is the guy lining up at center.
In the WHL, we use FOW = 0.01. (everything weighted compared to G=1, A=1). So a 1000 FOW center like Bergeron or Giroux or Crosby is worth an extra 10pts in a season. That "feels" about right.
[I absolutely loathe the combination of FOW with Yahoo!s player-eligibility-excessiveness... it so skews player value all over the place. One of the reasons I finally jumped over to Fantrax this year: positional OVERRIDE!!!]

Your G=18, A=18 would give me a headache!
I understand the optimism of "getting a lot of points", but when I watch a "Live Match-Up", I like to easily be able to do the math "Oh, I'm down by 0.9pts, I just need somebody to score a goal or assist!"

It's good to hear some people souring on PIM.
It really is a kind of awful stat that we've all started viewing as "GOOD" because of fantasy hockey tracking it as one of their more historical non-scoring stats.
But I loved when the Dobber leagues dropped PIM in favor of HIT... that was good.

+/-... I'm in the absolute minority, but I love +/-.
Ideally, shot-differential would be a much better metric to capture skater value of offense-vs-defense opportunity differential.

Guys like Nicklas Lidstrom or Patrice Bergeron... a lot of their real-NHL value was from limiting goals against.
Unfortunately, fantasy hockey doesn't have a good stat to capture "This player doesn't allow goals against".
If there was a SOGA/60min (shots against per 60min), perhaps that would be my ideal defensive metric to replace +/-... but that's not there yet on any platforms I've seen.
[Statistically this could be done by setting some ZERO benchmark for shots per minute. i.e. Perhaps give a player +0.05 for every minute on ice, say +1.0 for a 20min game and then taking away -0.01 for every shot against while on. So if player plays 20min, on for 8 shots against, he's +0.2 = 1-0.8. Of course, this would be hated like +/-... people just cry when they see a negative value show up. So if a guy played 20min and was on for 11 shots against, he'd be -0.1... and people would cry about how the scoring systems is set-up. Rant: I swear too many kids grow up these days getting a participation ribbon for every damn thing they do and people can't stomach a published scoring result that clearly suggests "Hey, you were less than average in the contest".)

I could go on about +/- forever...
Those on this site know that I passionately believe it is a very quality stat.
Even in some of my own leagues (NHLFL Dynasty), I've had to concede on its use due to majority opinion against it.
I... shall refrain from my arguments as a PRO +/- person... I have... no support in this battle in fantasy hockey circles! ugh.

[keytolife]

Good thoughts.

From being on these boards, it was a year or two ago that I read somebody's interesting set-up regarding FOW.
Somebody had said that Fantrax allows some stats to be counted only from certain positions - like FOW can be counted only from Centers.
I thought this was a great idea, as a fantasy GM (i.e. "coach"), somebody should be allowed to get some points from FOW from a "guy taking a face-off", which is the guy lining up at center.
In the WHL, we use FOW = 0.01. (everything weighted compared to G=1, A=1). So a 1000 FOW center like Bergeron or Giroux or Crosby is worth an extra 10pts in a season. That "feels" about right.
[I absolutely loathe the combination of FOW with Yahoo!s player-eligibility-excessiveness... it so skews player value all over the place. One of the reasons I finally jumped over to Fantrax this year: positional OVERRIDE!!!]

Your G=18, A=18 would give me a headache!
I understand the optimism of "getting a lot of points", but when I watch a "Live Match-Up", I like to easily be able to do the math "Oh, I'm down by 0.9pts, I just need somebody to score a goal or assist!"

It's good to hear some people souring on PIM.
It really is a kind of awful stat that we've all started viewing as "GOOD" because of fantasy hockey tracking it as one of their more historical non-scoring stats.
But I loved when the Dobber leagues dropped PIM in favor of HIT... that was good.

+/-... I'm in the absolute minority, but I love +/-.
Ideally, shot-differential would be a much better metric to capture skater value of offense-vs-defense opportunity differential.

Guys like Nicklas Lidstrom or Patrice Bergeron... a lot of their real-NHL value was from limiting goals against.
Unfortunately, fantasy hockey doesn't have a good stat to capture "This player doesn't allow goals against".
If there was a SOGA/60min (shots against per 60min), perhaps that would be my ideal defensive metric to replace +/-... but that's not there yet on any platforms I've seen.
[Statistically this could be done by setting some ZERO benchmark for shots per minute. i.e. Perhaps give a player +0.05 for every minute on ice, say +1.0 for a 20min game and then taking away -0.01 for every shot against while on. So if player plays 20min, on for 8 shots against, he's +0.2 = 1-0.8. Of course, this would be hated like +/-... people just cry when they see a negative value show up. So if a guy played 20min and was on for 11 shots against, he'd be -0.1... and people would cry about how the scoring systems is set-up. Rant: I swear too many kids grow up these days getting a participation ribbon for every damn thing they do and people can't stomach a published scoring result that clearly suggests "Hey, you were less than average in the contest".)

I could go on about +/- forever...
Those on this site know that I passionately believe it is a very quality stat.
Even in some of my own leagues (NHLFL Dynasty), I've had to concede on its use due to majority opinion against it.
I... shall refrain from my arguments as a PRO +/- person... I have... no support in this battle in fantasy hockey circles! ugh.

Oh i would love shot differential that would be a real great idea.

The reason why i don´t like +/- is the fact that is a stat that does not weight a ton because the diffrence is pretty minor and it often leads to improved value for the top teams which is something i am not a fan of either.

The thing with our points per night system that would boggle you is, that you get very earsily used to thinking this and that guy gives you X points per night usually. It does not come down to goals and assists that incredibly much because of the periphals counting. But to adjust them not not be too strong but still give defensive D-men that are valueable in the real NHL value, while also being able to balance goalies stats, those odd numbers are what we ended up with.

We discussed it prior to founding the league that we want to have players have a similar value than in the real NHL and played arround with some good examples and how much points those types should get. While i most leagues for example a Ekblad would be counting as struggeling, in oursetup he still is a very good D, not as good as last season but the improved hits and bs keeps things in place.

[JohnnyFantasy]

I think FOW can actually be linked to a specific value, even beyond what just "feels rights" (no offense Pengwin). I've been working on a blog post about finding a mathematical relation between all of the stat categories to weight them for a points league (post isn't up yet, but I am hopeful to get it up tonight and will let you know when I do if you are curious). In my research, I found a paper by a few statistics students that looked at the value of faceoff differential. If I recall correctly, the conclusion was that anything happening 20 or so seconds after a specific event in a game falls away into irrelevant background noise, and so looking at all of the goals occurring within 20 seconds after a faceoff, they concluded that it too a positive faceoff differential of about 75 FOW in order to result in a goal, on average. From there, you can take you take your goal value and divide by 75 to get the value of a FOW (in keytolife's case, 0.24). That forms a reasonable basis for FOW, but as you can see those additional points would be fairly negligible in the scoring scheme (especially considering that you should also value FOL at -0.24). But, it could always theoretically break a tie in a close matchup.

I agree on PIMs. Now that Hits is widely available, I don't see any reason to use PIMs (unless we are actually awarding negative points, which I think we should, at least for non-fighting penalties). We need to start a movement to break the stat-category inertia that everyone has gotten used to. Down with PIMs! I'm also against using wins and losses for goalies. As Pengwin stated, you can't eliminate a team's influence on any stat, but win is really an egregious offender in this respect. Take a look at Dubnyk so far this year. Should he really have any fantasy value? His team is winning games in spite of (not because of) his play. Compared to last year, where his team was winning because of his play - and then his more individual-based goal stats (such as save %) were excellent and he should have had (and did have) fantasy value because of it.

Peng, I'll come to your defense on +/-. I know why everyone is down on it (basically because it is awarded inaccurately - up to 2/3 of the time, as I understand). But, it is a scoring-based category and in categories leagues, we need more of those to balance out the peripherals. In points leagues, it's even better since we can weight the value. Think about whatever value your setup would normally give to +/-. If the problem with the stat is that it is only correctly awarded 2/3 of the time, multiply that value by 2/3, and then you have a setup that (on average, over the course of time) should accurately reflect what +/- was trying to get at (contributions to a goal being scored). I do want to hang onto this category until something like SOGA/60 comes into fantasy, because guys like Bergeron, Toews, etc. don't get as much fantasy value as they should. Granted, this doesn't completely solve their undervaluing, but it helps.

[Pengwin7]

I think FOW can actually be linked to a specific value, even beyond what just "feels rights" (no offense Pengwin). ...they concluded that it too a positive faceoff differential of about 75 FOW in order to result in a goal, on average. From there, you can take you take your goal value and divide by 75 to get the value of a FOW...


Peng, I'll come to your defense on +/-. I know why everyone is down on it (basically because it is awarded inaccurately - up to 2/3 of the time, as I understand).

Kind of a strange post.
In the first part, you suggestively disagree that I don't have the "feel right" thing correct.
My league weights G=1,A=1, FOW=0.01.
1 / 75 = 0.013.
I'd say I nailed our FOW weighting based on your math... I'm pretty decent at pegging a value at what "feels right" [mathematically, trust me... I'm your guy for math+hockey stats]

The +/- argument... ha, trust me, it doesn't at all feel like you are defending my stance.
Saying that +/- is awarded inaccurately up to 2/3 of the time is definitely nowhere near my stance on it.
You are in the majority here - basically pegging it as a highly flawed stat.

The NHL is well balanced though, and stats such as PDO suggest that "luck" involved with goalies letting in goals that shouldn't have scored or stopping goals that may should have scored... eventually... balances.
And I'd also argue that while there are instances where a player takes a minus from bad luck or teammate play... they'll also get a plus somewhere from good luck or teammate play.
That balances too.

What is left with enough sample-size (GP) is that the truly "PLUS" players will eventually shake out relatively better than their teammates and the truly "MINUS" players will be relatively worse than their teammates.

Upweighting or downweighting stats by some amount based on some speculation of +/- accuracy should not be the consideration.
+/- is relative to a benchmark of 0.
Whereas all other accumulating-only stats are only going to add value to a player.

Weighting +/- is the last step in setting category values.
I'd even say that weighting it (as a commisioner, or as a league) is as simple as saying:
"OK, with our stat weightings, the top player is worth 150pts in our set-up. How much do I want to change that amount if he's +30 or +0 or -30?"
That's how you find the proper +/- weighting for a league.
(It has nothing to do with this speculated 2/3 inaccuracy - no offense)

[JohnnyFantasy]

I'm not saying that you're feeling wasn't close, I'm just trying to make the point that we can get there in a more precise manner.

Well, I guess I am saying +/- is a flawed stat then. You're right that if everything worked out, the erroneous pluses would be balanced out by the erroneous minuses then no weighting would be required. But, part of the flaw of the stat is the way it accumulates on one side. For example, Jason Garrison, Marc Methot and Nik Hjalmarsson were all in the top 25 of +/- last year, many would argue due to being dragged up by the offensive powerhouses around them (Methot especially, playing with Karlsson) - probably a lot more false pluses than false minutes. From a fantasy perspective, reducing the amount of points awarded for +/- limits the overvalue to these players, but yes, this is based on the idea of compensating for a flawed stat.

Sorry, I'm trying to support you but I can't go all the way to saying there aren't problems with +/-. I do note a correction to my earlier post though: +/- is only accurately awarded 2/3 of the time (and thus only inaccurately awarded 1/3 of the time), and so it is correct more often than it is incorrect.

[StuntMan12]

It always puzzled me why PIM is counted in a positive note when its a negative stat because penalties hurt your team. Hence why I tend to stick with all points leagues

[JohnnyFantasy]

I agree, PIMs are negative and should be valued as such (in points leagues). Here is my newly published post that I mentioned before: http://johnnyfantasy.com/2015/11/06/...points-system/

Pengwin, I'm sure you'll disagree with the +/- section, but I'd love any and all feedback!