VHFL is Broken. Here's Why! (Part 1)

[Spartan 4,402]

Hi folks, I’ve been putting this piece off for quite a while since it’s taken a decent amount of work, and we all know I’m lazy. However, VHFL is something I’m quite passionate about. Back in my freshman year of college, a couple years before I joined the VHL, I took a course that taught me some pretty nifty Excel skills. I decided to mess around mid-season in S73 to make a spreadsheet that could calculate VHFL scores so that anyone could calculate their fantasy team’s scores, or check on the score for any player in the league at any given time. It was initially designed just to be a fun project for me to test out my skills on, and maybe have it used by the general community once I made it nice enough.

 

As luck would have it, Twists, the then VHFL manager stepped down during playoffs and left the league. Suddenly there was a need for a spreadsheet that could calculate VHFL scores quickly, and my file was soon adopted as the official VHFL calculation/management spreadsheet. All player rankings, user rankings, group rankings/scoring, and team tracking has all occurred in that sheet. I’ve tried to keep modifying and upgrading that spreadsheet to make life as easy as possible for whoever the current manager is, but it’s also allowed me to keep an eye on the state of VHFL over the past 5-6 seasons. I’m not liking what I’m seeing.

 

Warning

I'll preface this piece by saying there are a lot of statistics and numbers explained in here. There are also a lot of words. If you don't want to read it, just skip to the very bottom where I've written out a TLDR. 

 

So What's the Issue?

 

Before we get too deep into the issue at hand, let’s quickly look at VHFL scoring, just to be familiar with it: 

 

It’s pretty standard, 2 points for a goal, 1 point for an assist, 0.2 for hits, 0.8 for shots blocked. At a glance, it makes sense. Goals are worth double an assist, points per hit are pretty low since some folks can get over 300 hits in a season, and shots blocked can be hard to come by, especially for forwards. Makes sense to make them more valuable, right?

 

Well, as time has gone on, and we’ve seen seasons of skater rankings come out, there’s a very clear trend. Defensemen are simply way too good in VHFL. Take a look at this table, where I’ve tracked positional averages of the top 190-ish players in each season (190 chosen since that’s around the end of user controlled players in each season who put up any sort of stats):

 

S79 Positional Averages			S78 Positional Averages			S77 Positional Averages			S76 Positional Averages
C	61.98695652		C	131.7755102		C	130.7137255		C	134.5555556
LW	58.81428571		LW	133.3517241		LW	122.4064516		LW	114.16875
RW	62.29090909		RW	125.369697		RW	120.7028571		RW	116.6648649
D	86.5038961		D	198.1835616		D	216.9731343		D	194.7685714

 

Looking at these tables, the disparity is quite clear. Consider that around 190 skaters are in each sample size. Even in S79, where the data is as of mid-season - about 32-33 games in, defensemen already have about 24 more points over any other skater position. Going back before S79, the gap is more pronounced. 65 point gap in S78, 86 points in S77, 60 points in S76. Sure, we’ve had some extraordinary defensemen in the league, such as Condor Adrienne in recent seasons, but those sort of defensemen should be outliers. Defensemen shouldn’t be this far ahead of forwards. So where is the problem starting?

 

Brace Yourself, a Shit Ton of Statistics Incoming

 

Before we dive into some of the graphs I have coming up, I’d like to briefly explain a few statistics I’m going to use. In order to see exactly why there was this point disparity between forwards and defensemen, I decided to create a “point share” statistic inspired by the win share statistic that’s used in hockey generally, which is some breakdown of player value to their team. Value of a cog in the machine and all that, it’s super complex and just the inspiration to what I did. Instead of calculating player value to their team, I broke down the point share of each VHFL scoring attribute - goals, assists, hits, and shots blocked - to their final VHFL score.

 

---

 

Let’s use the gorgeous Alex Letang as an example. So far in S79, through 32 games, he has a VHFL score of 133.2, good enough to be the top player in fantasy. He has 10 goals, 48 assists, 106 hits and 55 shots blocked. In order to figure out his Hit Point Share and Shots Blocked Share, I individually calculated the amount of points generated by his hits and shots blocked.

 

Hits: 106 hits -> 21.2 points

Shots Blocked: 55 SB -> 44 points

 

When calculating the point share, we simply divide the points per attribute by the overall score of the player:

 

Hit Point Share: 21.2 hit points/133.2 total points = 15.92%

SB Point Share: 44 SB points/133.2 total points = 33.03%

Combined Hit+SB Point Share: 48.95%

 

So this shows a pretty normal point share of these statistics, for Alex Letang at least. Around 50% of his points are coming from 2 out of 4 scoring stats, which isn’t anything to worry about. But that’s only for Letang, who’s had the benefit of allowing his scoring on the ice to carry his VHFL score.

 

---

 

Let’s shift over to the 3rd best VHFL player this season: Mikeal Keef, a defenseman on the openly tanking Helsinki Titans. His combined Hit+SB Point Share is 75.57%. Over 75% of his 131 points are coming from hits and shots blocked! He has 8 goals on the season which is pretty good for a defenseman, but also only 16 assists. For comparison, he has more VHFL points than Red Lite, who has 29 goals and 33 assists. That just doesn’t make sense to me. The graphs even show the issue in scoring.

 

S79 Data (through 32-33 games)

 

I’ll be the first to admit that I should have broken the graph up into forwards and defensemen to make my point a bit more obvious, but I’ll just hope you take my word for how I break it down. I calculated the Hit Point Share and SB Point Share for every skater in the 190-200 sample from each season, and then graphed it out to see the distributions on a larger scale. There are two obvious clumps of red, which are the SB Point Shares. The lower group, which largely falls under 25%, are just about all forwards. The sole red dot under 25% between 120-140 VHFL points belongs to the aforementioned Red Lite. The entire clump that largely falls around or above 50% belong to defensemen. In the 120-140 range, there are 6 defensemen who are getting a significant chunk of their points from shots blocked. That’s a massive problem in my opinion, and the trend continues in all the seasons I observed. Take a glance:

 

S78 Data

 

S77 Data

 

S76 Data

 

In every single one of these seasons, there is a significantly large chunk of red hovering above 50% made up of defensemen, leeching a majority of their points from shots blocked. There’s also a bit of a parallel observation to be made here that forwards simply aren’t getting enough value either, outside the 2 or 3 outlier forwards that exist every season. The Boulet candidates for example who have massive amounts of hits, such as Groovy Dood, Christian Mingle, Mikko Lahtinen, and then the best players with checking on the worst teams who get a nice boost in shots blocked. It’s a very clear trend and it’s making VHFL too predictable. Don’t believe me? Let’s show you how the current scoring is creating a clear-meta in VHFL drafting that’s ruining the game.

 

Is There Really a Clear Meta?

 

In order to see if there’s a meta that users follow to win VHFL, I went back through the top users of S79 (as of 8/8) through S76 and pulled the top 5 in each season. I then went to their groups for drafting and noted down what positions they selected with their first, second and third picks of the draft. 

 

Here’s that breakdown:

 

S79 Top Users	First pick	Second pick	Third Pick
Doomsday (filler)	Vaakanainen - F	Letang - D	Nilsson - D
Fishy	Dood - F	Odinsson - F	Lite - F
Leafsman	Dood - F	Letang - D	Campbell - F
Tate	Lite - F	Letang - D	Lamb - F
Esso2264	Dood - F	Lamb - F	Pines - D
S78 Top Users	First pick	Second pick	Third Pick
Thunfish	Mingle - F	Kankkunen - D	Dear - F
Juice	Sandstrom - F	Mingle - F	Winter - F
Dasboot	Mingle - F	Sandstrom - F	Hornet - F
Renomitsu	Sandstrom - F	Elsby - D	Tsujimoto - F
GustavMattias	Hornet - F	Tsujimoto - F	Winter - F
S77 Top Users	First pick	Second pick	Third Pick
GustavMattias	Lahtinen - F	Dood - F	Walker - D
Victor	Killinger - D	Brown - F	Dood- F
Kylrad	Lahtinen - F	Killinger - D	Brown - F
gorlab	Letang - D	Dood - F	Marsh - D
leafsman	Lahtinen - F	Dood - F	Killinger - D
S76 Top Users	First pick	Second pick	Third Pick
OrbitingDeath	Lahtinen - F	Graves - F	Mitchell - D
Dasboot	Lahtinen - F	Mitchell - D	Bob - F
Juice	Lahtinen - F	Mitchell - D	Krishna - G
Esso2264	Dood - F	Hylands - F	Davis - D
Patrik Tallinder	Vikingstad - F	Hylands - F	Davis - D

 

Let’s all point and laugh at the poor folks in Doomsday's filler group. Must suck to suck. But charting out the picks made by each user and in each round, you get this table:

 

	Round 1	Round 2	Round 3
F	18	12	11
D	2	8	8
G	0	0	1

 

In the first round, 18 out of 20 users picked a forward first. The two exceptions were in S77, where Erik Killinger and Alex Letang were selected first, and they were 2nd and 4th overall in that season respectively - so really not a bad decision if you’re confident in landing a top 5 player, and also confident in your ability to find a top forward with your next few picks. Keep in mind, you have to select 3 forwards and 2 defensemen though. In all other situations where forwards were selected first, the forwards selected by these teams ended up top 3 in VHFL forwards. There’s a significantly higher emphasis on selecting a top forward with your first pick, as high end forwards are difficult to come by. There are usually only 2 or 3 forwards that land in the top 20 of overall skaters, and our top 20 users have prioritized those forwards with the first pick of their drafts.

 

Exactly half of them (10) also went back to selecting a forward with their 2nd pick as well. Common names here have been Groovy Dood in S77, Chris Hylands in S76, and Onde Sandstrom/Taro Tsujimoto in S78. The third round pick is usually one of the more crucial picks, and also similar to round 2 where most people go back to selecting a forward. Top scorers such as Robin Winter in S78, Grooby Dood in S77, Jim Bob in S76 have all been selected with the third pick of a user. One person, Juice, took Ajay Krishna with his third pick in S76, but also had Latrell Mitchell and Mikko Lahtinen with his first two picks to bail out that decision :P. Krishna was a top goalie that season anyways, so I suppose we can let it slide. 

 

The emphasis on forwards is undeniably clear. Out of the 60 picks made by the top 20 users in the past 4 seasons, over 66% of those picks are forwards. People know that they need to get the best forwards possible if they want a chance to win in fantasy. Defensemen are, as I like to say in Moscow, “a dime a dozen” simply because there are always 20-25 defensemen who will outscore the 5th best forward - VHFL-wise. The people who do select defensemen in the 2nd or 3rd rounds get it right though. They consistently selected defensemen who landed top 5 in the final rankings. Doomsday’s filler so far in S79 has both the #1 and #2 skater, in Letang and Nilsson. That kind of power can make up for a weak 3rd forward. And Vaakanainen, his 1st overall pick, was certainly the weak 3rd forward in the group as he went on to select Red Lite, the top forward in VHFL right now. But again, the argument here is that Nilsson shouldn’t be the 2nd best VHFL player just because he’s on a horrible defensive team that gives up plenty of shots against, that he can just stand in the way of.

 

---

 

This piece is already around 2,000 words, so I’ll save my “solutions” article for another day. I’ve thought of a few different ways to fix fantasy, but I’d love to hear suggestions from the community in case someone can think of something that I haven't. I’m happy to run testing on any kind of idea!

 

---

 

TLDR for you bums who skipped to the end: VHFL is broken because scoring favors the best defensemen on bad/average teams who block shots. Defensemen as a whole score way too much, with even average defensemen scoring up to 60-70 more fantasy points than the average forward. Therefore, VHFL success depends on how well a user can follow the meta of selecting the best forwards with their first few picks, and simply picking passable defensemen and an average goalie to finish out their team, since the top 20 skaters in every season have like 17 defensemen in them anyways. I think it's an issue, and will propose solutions to fix this down the road. If you have any thoughts, please comment them .

 

(2.4k+ words, 4 weeks claim)

 

Edited August 8, 2021 by Spartan

[Gustav 6,407]

Here's the thing--STHS favors certain players in certain ways. In general, there are a few things that jump out to me when I make my picks that signify possible top performance...here are some that come to mind.

 

-High CK rating. This is something we see people complaining about a lot, and two ways to address it could be to either negate the impact of hits entirely (which I think is boring) or to adjust the scoring so hits don't count for as much or they're offset by PIM counting for some sort of negative. I don't have a problem with physical players being rewarded more, and in fact I think it's an incentive for some people to choose to build a player with some CK--one of very few real build variations that's still considered "acceptable" to the big meta crowd.

-Good (or even just OK) players on bad teams--especially defensemen. My steal last season was Moscow's Bob Tristan, who had decent CK and the highest TPE among defenders on a rebuilding team. He ended up doing great. Last season's top scorer was Robin Galante Nilsson, who wasn't even all that impressive physically but who had a ton of SB and points for Vancouver. The thing is that it will always be meta to some extent to go for these players--I will concede that maybe SB are worth a bit much (almost an assist!), but outside of that potentially benefiting from being lowered, there's not much we can do about it.

-Anyone who's way better than the players around them. Groovy Dood was one of the hot picks of the past few drafts because he was often the only forward on DC even near his TPE level. Taro was that last season, and Tyler Walker was for one or two before that. I don't think there's anything that can be done to make those picks any less meta than they are unless we try to throw some stupidly complicated factoring-in-the-team thing into the equation, which I don't think is realistic and which I'm not a fan of anyway.

 

The only thing I would outright support is SB being weighted less heavily. I can understand the reasons why people complain about hits as well, but I'm more neutral on that.

[N0HBDY 1,119]

Us TLDR folks 

Edited August 8, 2021 by N0HBDY

[Spartan 4,402]

9 minutes ago, GustavMattias said:

Here's the thing--STHS favors certain players in certain ways. In general, there are a few things that jump out to me when I make my picks that signify possible top performance...here are some that come to mind.

 

-High CK rating. This is something we see people complaining about a lot, and two ways to address it could be to either negate the impact of hits entirely (which I think is boring) or to adjust the scoring so hits don't count for as much or they're offset by PIM counting for some sort of negative. I don't have a problem with physical players being rewarded more, and in fact I think it's an incentive for some people to choose to build a player with some CK--one of very few real build variations that's still considered "acceptable" to the big meta crowd.

-Good (or even just OK) players on bad teams--especially defensemen. My steal last season was Moscow's Bob Tristan, who had decent CK and the highest TPE among defenders on a rebuilding team. He ended up doing great. Last season's top scorer was Robin Galante Nilsson, who wasn't even all that impressive physically but who had a ton of SB and points for Vancouver. The thing is that it will always be meta to some extent to go for these players--I will concede that maybe SB are worth a bit much (almost an assist!), but outside of that potentially benefiting from being lowered, there's not much we can do about it.

-Anyone who's way better than the players around them. Groovy Dood was one of the hot picks of the past few drafts because he was often the only forward on DC even near his TPE level. Taro was that last season, and Tyler Walker was for one or two before that. I don't think there's anything that can be done to make those picks any less meta than they are unless we try to throw some stupidly complicated factoring-in-the-team thing into the equation, which I don't think is realistic and which I'm not a fan of anyway.

 

The only thing I would outright support is SB being weighted less heavily. I can understand the reasons why people complain about hits as well, but I'm more neutral on that.

Most of what you've said is about checking, which I've never stated was an issue. If anything, one of my solutions that I was working on would actually reduce the value of shots blocked and slightly increase the value of hits since it's a bit more controllable. In the sense that you choose how much you want to hit, but shots blocked is mainly just a result of team state.

 

Tristan and Nilsson are two sides of the same coin. Nilsson's number is widely overinflated because of shots blocked. Checking is fine, Nilsson didn't check much but 75% of his points coming from SB is an issue. Tristan also was the best defenseman on his team, which isn't an issue in a vacuum, but considering that he was like 450 TPE last season and 53% of his points came from blocked shots and 20% from hits, I'd say that's a bit more of an issue. 

 

And I don't have any issue with the Taro, Adrienne, Walker, Dood picks either, they're ideally the ones who would have strong fantasy seasons because of being the best players on bad teams. I was also thinking of slightly punishing players on bad teams since they get a free chance to stat pad, perhaps by using +/- as a factor in some linear boost/nerf. Like +/- multiplied by 0.2 or something of that sorts. Good players on good teams usually see less fantasy points just because STHS spreads out the love. They can only be top fantasy options if they play like Lahtinen/Letang/Rice/Davis on S76 Moscow and all record a shit ton of points, therefore a lot of fantasy love.

 

Overall though, yes, reducing shots blocked is the priority. I would also like to reward scoring a bit more, since the actual points scored should be a true measure of a player first and foremost. Physicality statistics should be secondary.

Edited August 8, 2021 by Spartan

[N0HBDY 1,119]

39 minutes ago, Spartan said:

Most of what you've said is about checking, which I've never stated was an issue. If anything, one of my solutions that I was working on would actually reduce the value of shots blocked and slightly increase the value of hits since it's a bit more controllable. In the sense that you choose how much you want to hit, but shots blocked is mainly just a result of team state.

 

Tristan and Nilsson are two sides of the same coin. Nilsson's number is widely overinflated because of shots blocked. Checking is fine, Nilsson didn't check much but 75% of his points coming from SB is an issue. Tristan also was the best defenseman on his team, which isn't an issue in a vacuum, but considering that he was like 450 TPE last season and 53% of his points came from blocked shots and 20% from hits, I'd say that's a bit more of an issue. 

 

And I don't have any issue with the Taro, Adrienne, Walker, Dood picks either, they're ideally the ones who would have strong fantasy seasons because of being the best players on bad teams. I was also thinking of slightly punishing players on bad teams since they get a free chance to stat pad, perhaps by using +/- as a factor in some linear boost/nerf. Like +/- multiplied by 0.2 or something of that sorts. Good players on good teams usually see less fantasy points just because STHS spreads out the love. They can only be top fantasy options if they play like Lahtinen/Letang/Rice/Davis on S76 Moscow and all record a shit ton of points, therefore a lot of fantasy love.

 

Overall though, yes, reducing shots blocked is the priority. I would also like to reward scoring a bit more, since the actual points scored should be a true measure of a player first and foremost. Physicality statistics should be secondary.

TLDR?

[Spartan 4,402]

15 minutes ago, N0HBDY said:

TLDR?

 

monke

[CowboyinAmerica 2,889]

I absolutely agree with you that defensemen are way more powerful in VHFL, and I've personally used the "focus on shots blocked" strategy to good effect in past seasons. But the two points I'd say as devil's advocate:

 

1. Try and predict the shots blocked leaders. It's easier said than done - I know because I've tried. Generally you're looking for players that have a high skating/defense skill and are on bad teams that have a lot of shots made against them. But it's not a perfect 1:1 measure by any means, and if you miss, you miss by a lot since those players don't have the goals/assists to provide a buffer in their scores. It's a boom or bust strategy essentially - it doesn't surprise me that a lot of your top scorers are chasing that strategy, but I'd bet the more knowledgeable players who still score poorly are doing the same. And I don't mind shots blocked being a high score with that variability.

 

2. I'd be curious whether that F first then D meta you've identified is actually different from the mean in any way. Historically, I feel like forwards are 5/6 of the first round anyway, since they've tended to be the top goals/assists getters and that's what most people tend towards anyway. I'm not sure it's as much what the best players are doing as much as it's what everyone is doing.

[Victor 10,940]

Agreed. Relying on a scoring system I came up with in S36 or something might be outdated.  

[Garsh 1,190]

2 hours ago, Spartan said:

Overall though, yes, reducing shots blocked is the priority. I would also like to reward scoring a bit more, since the actual points scored should be a true measure of a player first and foremost. Physicality statistics should be secondary.

 

Sorry defensive defensemen

[Spartan 4,402]

37 minutes ago, CowboyinAmerica said:

I absolutely agree with you that defensemen are way more powerful in VHFL, and I've personally used the "focus on shots blocked" strategy to good effect in past seasons. But the two points I'd say as devil's advocate:

 

1. Try and predict the shots blocked leaders. It's easier said than done - I know because I've tried. Generally you're looking for players that have a high skating/defense skill and are on bad teams that have a lot of shots made against them. But it's not a perfect 1:1 measure by any means, and if you miss, you miss by a lot since those players don't have the goals/assists to provide a buffer in their scores. It's a boom or bust strategy essentially - it doesn't surprise me that a lot of your top scorers are chasing that strategy, but I'd bet the more knowledgeable players who still score poorly are doing the same. And I don't mind shots blocked being a high score with that variability.

 

2. I'd be curious whether that F first then D meta you've identified is actually different from the mean in any way. Historically, I feel like forwards are 5/6 of the first round anyway, since they've tended to be the top goals/assists getters and that's what most people tend towards anyway. I'm not sure it's as much what the best players are doing as much as it's what everyone is doing.

For point 1, I think that trying to predict shots blocked leaders are part of the strategy, at least for defensemen. Shots blocked leaders are generally always defensemen, and you know that if you select a top shot blocker, he will be a top 30 VHFL player. Take Cowboy Prout for example, he was someone I selected nearly every season before Calgary became better, and he'd consistently be a pretty decent defenseman and land up top 30 simply because he was a high TPA D-man on a non-playoff team. 

Out of amusement though, here's Cowboy Prout's ranking and SB Point Share over the past 4 seasons:

S79 - 45th skater, 43% SB Point Share

S78 - 5th skater, 55% SB Point Share

S77 - 21st skater, 59% SB Point Share

S76 - 38th skater, 57% SB Point Share

 

He's had a lot of variability, as is to be expected from any player who isn't offensively driven, but when 55% of his points are consistently coming from Shots Blocked, that's when I have an issue with it. It's some kind of happy medium that we need between scoring and the physical stats to provide more balance and variety. 

 

As per your second point, I don't think that this will differ from the mean in general, although I do see a lot of people who pick defense early and often lose their groups consistently. The only exception is when you manage to target a top 5 defenseman appropriately. Sorry to all the folks who picked Roque Davis in the 1st round after his trade to LA . But also you prove my point that everyone targets forwards early, as that's a strategy that really shouldn't be a thing. Following a blueprint is boring, and it just results in everyone doing the same thing every season. No one ever picks a goalie first either, because goalie scoring ends up having most of the starting goalies in a 20 point range from each other. Good forwards shouldn't be that scarce to find, which is also why I've been debating of some way to reduce the point disparity between forwards and defensemen to make selecting defensemen in the first round viable as well.

 

[Spartan 4,402]

3 minutes ago, Garsh said:

Sorry defensive defensemen

Yeah, this is something that will definitely suck a little bit. But NHL Fantasy ends up comparing a teams stats to each other and then seeing which team wins out per statistic. Like a point to the fantasy team with more goals. We just have a different way of running fantasy, which is also why defensive defensemen have never really been viable in VHFL. They may make "defensive" builds, but they will inevitably score points when they're on the average/below average teams in the league.

 

I'd just like to stop seeing this every season:

 

And honestly, 2 forwards in the top 15 has only happened one other time in the last 4 seasons, back in S76. There were no forwards in the top 15 in S78 or S77. 

[Gustav 6,407]

6 hours ago, CowboyinAmerica said:

I'd be curious whether that F first then D meta you've identified is actually different from the mean in any way. Historically, I feel like forwards are 5/6 of the first round anyway, since they've tended to be the top goals/assists getters and that's what most people tend towards anyway. I'm not sure it's as much what the best players are doing as much as it's what everyone is doing.

 

I consciously pick forwards first because they're more obviously predictable to the casual player. Someone I really want at forward will probably be off the board if I let them fall, but knowing how the engine treats people I know I can find some 500-TPE defenseman on a depleted roster later on who will do great things for my team.

 

I think changing the system to combat meta might limit that to some extent, but those players are still favored pretty well by the engine so it will probably always be meta to some extent regardless of scoring system.

[Prout 373]

6 hours ago, Spartan said:

For point 1, I think that trying to predict shots blocked leaders are part of the strategy, at least for defensemen. Shots blocked leaders are generally always defensemen, and you know that if you select a top shot blocker, he will be a top 30 VHFL player. Take Cowboy Prout for example, he was someone I selected nearly every season before Calgary became better, and he'd consistently be a pretty decent defenseman and land up top 30 simply because he was a high TPA D-man on a non-playoff team. 

Out of amusement though, here's Cowboy Prout's ranking and SB Point Share over the past 4 seasons:

S79 - 45th skater, 43% SB Point Share

S78 - 5th skater, 55% SB Point Share

S77 - 21st skater, 59% SB Point Share

S76 - 38th skater, 57% SB Point Share

 

He's had a lot of variability, as is to be expected from any player who isn't offensively driven, but when 55% of his points are consistently coming from Shots Blocked, that's when I have an issue with it. It's some kind of happy medium that we need between scoring and the physical stats to provide more balance and variety. 

 

As per your second point, I don't think that this will differ from the mean in general, although I do see a lot of people who pick defense early and often lose their groups consistently. The only exception is when you manage to target a top 5 defenseman appropriately. Sorry to all the folks who picked Roque Davis in the 1st round after his trade to LA . But also you prove my point that everyone targets forwards early, as that's a strategy that really shouldn't be a thing. Following a blueprint is boring, and it just results in everyone doing the same thing every season. No one ever picks a goalie first either, because goalie scoring ends up having most of the starting goalies in a 20 point range from each other. Good forwards shouldn't be that scarce to find, which is also why I've been debating of some way to reduce the point disparity between forwards and defensemen to make selecting defensemen in the first round viable as well.

 

 

Cowboy Prout goat.

 

 I already hit my peak tho 

[dasboot 521]

TLDR I’m too good at vhfl 

[OrbitingDeath 3,183]

I'm here to read the tldr and leave a like

[Mr_Hatter 1,608]

First off, great article @Spartan! Obviously a lot of work went into this. My only major critique (without diving into the stats yet, since I have to work) is that I disagree with the sentiment that defensemen are overvalued, for one specific reason: its all relative. Due to the roster constraints, everyone picks the same number of people at any given position. So saying defensemen are better than forwards at getting fantasy points is a non-issue, as its not like you can pick 3 defensemen. It would be like saying forwards score more than goalies; the position is different, you can only pick an exact amount of each, so the differences in points between positions is immaterial. The only reason it feels worse for dmen-> forward as opposed to forward-> goalie is that dmen and forwards calculate using the same scale. But the actual number itself doesn't matter. 

 

That being said, I think the major takeaways are points percentages coming from certain stats, which is a fair one. I don't 100% agree that it's always a problem that good players on bad teams will perform better in fantasy due to higher SB numbers,  as this whole VHFL game is a meta game to begin with. Learning how to pick players is part of that, and an important part of people's enjoyment.

 

I think you could argue the numbers could be rebalanced, but I'd be wary of going to far in the wrong direction. If we completely nerf SB then there is no reason to pick players from bad teams, which completely destroys the fun of the game for me. You should be able to consider players from all teams, so there has to be some benefit to being on a bad team, which is usually what that SB total represents. 

[Mr_Hatter 1,608]

Followup/counterpoint to my original response; the differences between F and D aren't important IMO, but the differences between D and D are.

 

IE, we should try and maximize the points differential between top F and low F, top D and low D, and top G and low G, not the differential between top F and top D. That makes the game more exciting, because then every pick matters. So we want the standard deviation from F1 to F30 to be similar in scale to D1 to D20, to G1 to G10. That way, you prioritizing certain positions at certain times becomes more important. 

 

Idk just some thoughts. I might be wrong haha

[Spartan 4,402]

8 minutes ago, Mr_Hatter said:

My only major critique (without diving into the stats yet, since I have to work) is that I disagree with the sentiment that defensemen are overvalued, for one specific reason: its all relative. Due to the roster constraints, everyone picks the same number of people at any given position. So saying defensemen are better than forwards at getting fantasy points is a non-issue, as its not like you can pick 3 defensemen. It would be like saying forwards score more than goalies; the position is different, you can only pick an exact amount of each, so the differences in points between positions is immaterial. The only reason it feels worse for dmen-> forward as opposed to forward-> goalie is that dmen and forwards calculate using the same scale. But the actual number itself doesn't matter. 

This is a fair point, that everyone is experiencing the same "issue." However, I found this case more interesting because it overall devalues defensemen when it comes to making fantasy selections. You'd be fine with D20, since the gap from D1 to D20 isn't that large. Both players will still be in the top 30 of overall skaters. However, the difference between F1 and F20 is significantly large, from top 30 to maybe top 50. Defensemen simply are too easy to coast on, there isn't really much of a skill component to selecting them. You pick defensemen on bad teams who will get shot blocks, and defensemen on high scoring teams who will get fantasy points from scoring. But picking the top 2 defensemen on the bottom 12 teams in the league is much more prevalent in VHFL strategy, in my opinion.

 

13 minutes ago, Mr_Hatter said:

goalies

So this is straying away from what I've written so far, but I do plan on looking at goalies as well to create a bit more of a disparity as you had mentioned, to try and make the gap from G1 to G12 a bit wider. 

 

 

This is the current state of goalies, a 12 point gap from G1 to G10. If you pick literally any of these, your goalie position won't be the reason you lose. Unless you're one of the rare unlucky folks who lose out on a position by <5 points. RIP to those fine folks. But yeah, making this gap wider is something I'd like to think about.

 

20 minutes ago, Mr_Hatter said:

I don't 100% agree that it's always a problem that good players on bad teams will perform better in fantasy due to higher SB numbers,  as this whole VHFL game is a meta game to begin with. Learning how to pick players is part of that, and an important part of people's enjoyment.

 

I think you could argue the numbers could be rebalanced, but I'd be wary of going to far in the wrong direction. If we completely nerf SB then there is no reason to pick players from bad teams, which completely destroys the fun of the game for me. You should be able to consider players from all teams, so there has to be some benefit to being on a bad team, which is usually what that SB total represents. 

This is probably going to be the major argument against making any changes to VHFL - it's just something for fun, doesn't need to be a super overthought process. It's mainly out of my own impulse that I decided to look into why VHFL felt too cookie-cutter, and I found what I presented in the article. However, STHS still favors the best players on bad teams just from a points standpoint. They'll still be very viable, I also don't want to eliminate those players from viability. I just don't want those players on average/bad teams to be dominating VHFL, and some more variability would be really nice. 

 

35 minutes ago, Mr_Hatter said:

So saying defensemen are better than forwards at getting fantasy points is a non-issue, as its not like you can pick 3 defensemen.

I think my issue with this is also that your 2nd and 3rd forward combined may not score as much as one of your defensemen. If you have 2 top 10/15 forwards, an average goalie and any defensemen who doesn't fall out of the top 30 for overall skaters, you will land at least 3rd, likely 1st or 2nd with those forwards. The strategy is mostly clear, as a few others have pointed out -> you prioritize selecting the best possible forwards and leave defense for later since you're more likely to get a defensemen in the top 30 of skaters in the 4th or 5th round than you are to get a forward in that same top 30.

 

34 minutes ago, Mr_Hatter said:

IE, we should try and maximize the points differential between top F and low F, top D and low D, and top G and low G, not the differential between top F and top D. That makes the game more exciting, because then every pick matters. So we want the standard deviation from F1 to F30 to be similar in scale to D1 to D20, to G1 to G10. That way, you prioritizing certain positions at certain times becomes more important. 

 

In my perfect world, picking defensemen and goalies in the 1st round would be viable and encouraged. Users must try to figure out who would truly be the best player for their team, and no position is easier to pick between. As you mentioned, emphasizing and maximizing the disparity within position groups is key, and to do that, I think not only do we need to reduce the impact of Shots Blocked to reduce defensemen point inflation, but also increase the value of multiple other scoring stats to widen the gap between players of the same position. Making goals and assists more valuable, from 2 and 1 to 3 and 1.5. Rewarding hits more. Making shutouts significantly more valuable. Punishing for goals against. 

 

 

 

[Dil 1,760]

Alternate Title: STHS is broken. Here's Why!

[Mr_Hatter 1,608]

Thanks for the in depth response! I definitely see your perspective on a lot of stuff, and for sake of brevity I won't respond to everything

 

7 hours ago, Spartan said:

I think my issue with this is also that your 2nd and 3rd forward combined may not score as much as one of your defensemen. If you have 2 top 10/15 forwards, an average goalie and any defensemen who doesn't fall out of the top 30 for overall skaters, you will land at least 3rd, likely 1st or 2nd with those forwards. The strategy is mostly clear, as a few others have pointed out -> you prioritize selecting the best possible forwards and leave defense for later since you're more likely to get a defensemen in the top 30 of skaters in the 4th or 5th round than you are to get a forward in that same top 30.

This is the only thing I disagree with you on. Since we are limited to exactly 3F 2D 1G, defensemen earning more doesn't mean anything in terms of overall fantasy performance. IMO you are conflating the total points of F/D with the spread of F and D fantasy point distributions. Like, the issue here isn't that Dmen earn 100 points more on average or whatever, its that the drop off from higher performing F is much higher than the drop off from higher performing D.

 

If we were able to make the bell curves of F to D points roughly equal, in proportion with the number available/required to select at that position, then that would make picking Dmen earlier more viable, as your expected drop off relative to that position would be more equal, in terms of the standard deviation at that position. (I recognize that you agree about this in the quote later on, but for discussions sake I'm going to keep talking on this point)

 

Instead what we have is a class of F that has a few high earners before a significant drop off, whereas with D you have some high point getters sure, but there is an oversaturation of medium-high point getters relative to the number of Dmen that actually need to be selected. 

 

But that doesn't matter in regards to the balance between the point totals received by D/F. Like the best way to explain this is to look at the logical extremes of the situation. If there were 10 Dmen available to pick for 10 slots, and five scores 400 points with another 5 scoring 350, you don't need to prioritize D. Whereas if there are 15 F for 15 slots, and the top 5 are at 300, while the rest are at 150, obviously you want to pick F first, because otherwise you will miss out on one of the good forwards. But if we were able to improve it where, generally speaking, those 10 D are spaced between 200 and 400 in 20 point increments, and the 15 F are between 150 and 300 in 10 point increments, there is more room to pick D early, depending on how you read their team situation/build. 

 

idk I'm not sure that my math is 100% correct in this exact scenario, but it gets my point across. The more you balance the point distributions within a position relative to the number of pickable players, the more viable picking any position at any time is. 

 

7 hours ago, Spartan said:

 

In my perfect world, picking defensemen and goalies in the 1st round would be viable and encouraged. Users must try to figure out who would truly be the best player for their team, and no position is easier to pick between. As you mentioned, emphasizing and maximizing the disparity within position groups is key, and to do that, I think not only do we need to reduce the impact of Shots Blocked to reduce defensemen point inflation, but also increase the value of multiple other scoring stats to widen the gap between players of the same position. Making goals and assists more valuable, from 2 and 1 to 3 and 1.5. Rewarding hits more. Making shutouts significantly more valuable. Punishing for goals against. 

Yeah I probably would be down for lowering SB's a little, but I don't agree with raising hits. They are already overvalued IMO. I think I probably like raising goals and assists a little. I'd like to explore kinds of team modifiers that might raise or lower your score in some ways, I'm just not sure of how to do that effectively/fairly. As far as goalies, I am hesitant to increase shutouts, as they are incredibly random/team dependent. I think we need to factor in some kind of penalty for GA though, or just generally some way of factoring in SV% as opposed to only considering base number of saves, as I think it slightly overvalues goalies on bad teams, regardless of that goalies performances. 

 

[Spartan 4,402]

Claim 2/4

[Spartan 4,402]

Claim 3/4

[Spartan 4,402]

Claim 4/4

[a_Ferk 766]

Basically a good summary of all of my issues with VHFL as well. My biggest complaint with it was that goalies don't really matter a lot. The difference between the best goalie and the worst one is minimal, and I think that it could be changed as well. Checking/defensive defensemen are also completely broke as you said, and really need to be nerfed. 10/10

[Spartan 4,402]

On 9/13/2021 at 4:28 PM, a_Ferk said:

Basically a good summary of all of my issues with VHFL as well. My biggest complaint with it was that goalies don't really matter a lot. The difference between the best goalie and the worst one is minimal, and I think that it could be changed as well. Checking/defensive defensemen are also completely broke as you said, and really need to be nerfed. 10/10

Yep, I had only really focused on skaters, but I do recognize that goalies need more differentiation as well. I haven't explored avenues for that yet, but it will certainly be on the docket. Thanks