Fantasy Hockey Geek Ranking Question

[Yanni72]

Hi, I'm wondering if I am doing something wrong. My league is:

Yahoo H2H Roto.

2C, 2R, 2L, 4D, 2G, 2 Util, 4 bench

G, A, +/-, PPP, GWG, SOG, Hits, W, L, GAA, Save %, SO

Based on Dobber's latest projections:

Is it possible that Byfuglien would be ranked 6th!? After Karlsson and before Letang?

Phaneuf ranked 12th!? Ahead of Neal & Tavares?

TIA

[Pengwin7]

FHG's ranking are solid.

1. The HITS category may propel Buff slightly ahead of Karlsson. Buff had 112 hits last year vs. Karlsson's 60. That's almost double. There's not too many categories where Karlsson is going to double Buff... maybe assists. Everything else can be fairly equal. Think of it that way. Buff also will have an edge in SOG, if healthy.

2. Phaneuf (& defensemen in general) will be ranked higher based on the roster allocation. There are 6F spots, so a 12-team pool has 72F in play. But how much more valuable is Tavares or Neal than the 73rd best forward? There are 4D spots, so a 12-team pool has 48D in play. It's possible that Phaneuf is just maybe 12.8pts higher than the #49D and maybe Tavares or Neal are 12.5pts higher than the #73F. That's all.


What matters is how you use this information to do a draft.
You want to take a guy as late as you can... so if you can Phaneuf in the 4th round, you do it.

I took Tavares 6th overall in two multi-cat pools this year. Is he the 6th most valuable fantasy-player, no... probably not, but he will be very solid & dependable and he won't be around in the 2nd round, while the D-men will still be there.


Anyways, you have to mix what you learn from FHG rankings with knowing how your leaguemates will draft. Just try to draft those high FHG guys before others do... but remember that pools are won on "value"... so the later you take them, the more "value" there is to your team overall.

[Yanni72]

Thanks for the reply Pengwin. Can you please rephrase why defensemen in general are ranked higher? I have trouble understanding.

[Pengwin7]

I have my own spreadsheet showing player value (broken down by category) in a league that includes Hits/Blocked Shots.

For defensemen:
The elite defensemen have a value of 6.00+.
Then there is a tier of about five D that have a value of 5.00+
And then about 15 D-men are in the 4.00+ range.
But the scores quickly (by #25!) slip below 4.00.

For forwards:
Look, the forwards don't slip below a value of 4.00 until forward #91.


So... let's talk about player values above 4.00:
90 forwards do it
25 defensemen do it


Your league lists 8F (+Util?)... and forward #96 (12x8) will still get your over 4.00pts of fantasy value.
Your league lists 4D (+Util?)... and defense #48 (12x4) will only get you about 3.30pts of fantasy value.

So... a solid forward of 5.00pt value is +1.00 above that "fringe" forward, taken at the end of your draft.
And... a defenseman of 5.00pt value is +1.70 above that "fringe" defenseman, taken at the end of your draft.

Both the forward & defenseman get the same points... but it is this "relative difference" that gives the defenseman more value.


To explain it one more time, more simplistically, consider your 3rd round pick & 12th round pick.
Which is more points:
3rd round forward (70pts) + 12th round defense (25pts)
or
3rd round defense (50pts) + 12th round forward (55pts)


Clearer?

[Yanni72]

Much clearer. Thanks Pengwin.

[Hey Robbie]

I've always used and supported the FHG (HPG) rankings, and for the reasons well explained by Penguin here. However I have a comment or two for Yanni and, more importantly, a question that arose for me after reading the article "Understanding Your League" written by Terry Campkin on Wednesday.

1. While the rankings are more clear and probably a better direct guide to draft strategy in a points-only scenario as in Penguin's example, things get more complicated in multi-cat. The rankings, whether from FHG or Penguin, provide a combined value from the various cats. As such, they do a good job of explaining each player's value in isolation relative to the population of all other players (at least skaters; the normalization with goalies is more confounding for me), it does not necessarily mean that the best TEAM would be assembled by maximizing overall values. Penguin has also done an excellent job explaining in many posts the potential benefits of drafting for points earlier and hits/blocks/etc. later.

2. Drafting strategy also depends greatly on the other GMs' tendencies. If YOU know that Byfuglien may actually be more valuable than Tavares but your opponents do not, there is no sense drafting Buff in accordance with his actual value, but rather as informed by his perceived value by the other GMs.

3. The values of players from many tools based on projections often do not factor in the differing levels of confidence in projections of the various categories. This is notoriously a problem for certain cats, e.g. plus/minus and short-handed goals. If plus/minus is weighted the same as assists in an overall value calculation, it will be less useful than one that weights the categories versus others. N.B. this problem is ameliorated if the projections themselves incorporate year-to-year variance.

4. The value calculations presented by the model can only be as good as the projections used to calculate them. As with all models, garbage in, garbage out.

In general, all of the above speak to the fact that valuations by FHG and other quantitative mechanisms are not intended to be, nor she they be used as, draft guides dictating draft order. They are tools that, along with many other factors, can help you analyse the player pool and devise a draft strategy.

[Shoeless]

Originally Posted by Yanni72

Hi, I'm wondering if I am doing something wrong. My league is:

Yahoo H2H Roto.

2C, 2R, 2L, 4D, 2G, 2 Util, 4 bench

G, A, +/-, PPP, GWG, SOG, Hits, W, L, GAA, Save %, SO

Based on Dobber's latest projections:

Is it possible that Byfuglien would be ranked 6th!? After Karlsson and before Letang?

Phaneuf ranked 12th!? Ahead of Neal & Tavares?

TIA

Your set-up was right Yanni - I got the same results as you for skaters when I input your stuff in as a 12 team league - I used max games of 82 but there seems to be a glitch in the system where it also requires the same max games for bench positions so I am guessing that there might be some skew in results - I'm going to email the Geeks to let them know what I found.

Edit: Chatted with Paul and the issue of max games is a non-issue so yeah according to the FHG methodology, those rankings are accurate for your team.

[Hey Robbie]

Now for the question about FHG rankings and Campkin's article:
First a caveat that I fully expect FHG, Metaldude, Penguin, and a host of others with superior insight to mine to point out why the analysis I am about to give is specious, and I look forward to such criticism.

To look at the utility of FHG's rankings I ran a simulation comparing teams' production to their overall values, defined as the sum of the FHG values of all players on the roster. To simplify matters, I modeled based on a 10 team, points-only, roto league. Each team was assumed to have six forwards and four defensemen. I then used FHG to determine player values with the input being actual stats from last season.

I selected random teams from a pool of the rosterable forwards and defensemen (i.e. top 60 and top 40 respectively). Because the total available ranks in this pool is 5000, I defined as reasonably realistic teams as those with a sum of 495-505 total ranks, and eliminated all teams not meeting this criterion. I then made runs producing at least 5000 reasonable teams. For each team, I added the total number of points for the 10 players as a measure of actual performance.

I divided the reasonable teams into three classes: those with lower than expected average forward rank ("better" forwards), those with lower than expected average defenseman rank ("better defensemen"), and those more or less balanced. For each class I calculated the average number of points per team.

My preliminary results were that teams with more highly-ranked (i.e. with a lower than expected total number of ranks) forwards had a higher average production than balanced teams or teams with "better" defensemen.

I do not have my data in front of me right now, and know that many questions will require the particulars. Tonight I should be able to provide data and quantitative analyses, but I wanted to throw the methodology out there now in case I'm missing something obvious I can correct before going through the trouble.

If the preliminary results hold up, does this cast doubt on the common assertion, which I myself have frequently argued against on these boards, that it just doesn't make sense to rank above-average defensemen above average forwards because forwards score more points?

[fantasyhockeygeek]

First a caveat that I fully expect FHG, Metaldude, Penguin, and a host of others with superior insight to mine to point out why the analysis I am about to give is specious, and I look forward to such criticism

I thought I heard my ears burning!

If the preliminary results hold up, does this cast doubt on the common assertion, which I myself have frequently argued against on these boards, that it just doesn't make sense to rank above-average defensemen above average forwards because forwards score more points?

Nope, for a couple reasons.

The first is that you're making an assertion that it doesn't make sense to rank above-average D over average F. The value on FHG is actually "value over replacement", ie value as compared to the best option available on waivers. This "replacement player" varies greatly based on league size.

I'm not sure that basing things of "ranks" is fair -- by doing it that way, you ignore gaps in valuation ie you're assuming that the gap between player #1 and player #2 is the same as the gap between player #2 and player #3, aren't you? That, to me, is a huge misrepresentation of player value.

If you're looking at a population of 60F and 40D, you're looking at roughly the top-2 forwards on each NHL team and the top 1.33 D, right? Usually, D are pushed up in value by having more of them owned, because below the top 60ish you see a big drop-off in production.

I guess I'm having a hard time with your drawing a general conclusion like this based on one specific setup, when the correct answer is that the value of defense is dependent on the league size and scoring categories.

[lucifer316]

Originally Posted by fantasyhockeygeek

The value on FHG is actually "value over replacement", ie value as compared to the best option available on waivers. This "replacement player" varies greatly based on league size.

basically vbd right