Page 1 of 4 1 2 3 4 LastLast
Results 1 to 15 of 60

Thread: Average Numbers from Top Positional Players

  1. #1
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default Average Numbers from Top Positional Players

    Here's the excel file of the numbers from the Kessel thread.

    Basically I looked at the averages, gave it a green if it was a +5 more than the average, if it was a minimal difference < 5 and >-5, I gave it a yellow, and if was < -5 it was a red...

    It's generally believed that RW are much weaker than C and LW from an overall perspective (3rd LW is generally a better producer than a 3rd RW), but what's interesting is the variation between the tiers of RW aren't all that much.

    Eg.) If I went with Iggy (Yahoo! ADP of 37th) instead of Kessel (ADP of 18th), I'm essentially getting the same thing, but a round or two later...

    *A couple of downfalls of what I did, dual eligible players mucked up the averages a bit, so Malkin in C and RW could have skewed the numbers a bit. Also injured players, I took the totals from what they posted last season regardless of games missed, which once again could have skewed the numbers, Crosby's numbers would have brought the average down and made a player like Sedin look slightly better...

    Have at er!
    Attached Images Attached Images
    Dobber Expert Pool Champion 2011-12

  2. #2
    Rep Power
    17

    Dobber Sports Apprentice

    Default

    I'd suggest using z-scores rather than raw totals. Puts everything on the same scale. 5 PPP are much more significant than 5 SOG. It then becomes easy to just sum the z-score totals for each player to get an idea of the talent at the top of each position. You can also run a few iterations on the whole pool of players and find out who the best 14 are at each position rather than just using the top 14 drafted, which are almost always not the best 14 players.

    Z-scores are calculated as ("X" - Average)/Standard Deviation. It's a really simple calculation and gives much better insight into the numbers. Also, you can do conditional formatting and colour-code everything at once since they're all on the same scale.

    Generally all forward positions should be treated the same, especially if there is a utility spot. Replacement level for LW, RW, and C is almost certain to be the exact same. If your league mates are putting extra value on RW or something, I'd suggest to zig when they zag, and grab the elite guys at the so-called "deep" positions, and settle for the lower tiered players at "shallow" positions. If your'e in a really shallow league like 10 teams and only 6 forwards per team with no util spots, then you start to run into position scarcity, but I get the feeling most people doing this type of analysis are in deeper leagues like 12+ teams with 9+ forwards per team.

  3. #3
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default

    Players should always be placed at the weakest position for analysis.
    For this year, I've judged the following:

    C/LW is a LW
    C/RW is a RW
    LW/RW is a RW
    RW/D is a D

    So... THIS year, RWs are as strong as ever.
    When/if the statistical providers ever get their positional act together, then guys like Malkin, Giroux, and Backes should all be C only. Guys like Zetterberg, Datsyuk, Benn, M.Richards, Ott should also be C only (not C/LW).

    Here is a link to all dual Yahoo! guys for this year 2012-2013.

    Also, when I was tracking player-for-player, I used multiple counters to sum the total players at any forward position and also sum them in my preferred position (per above).
    That thread is here... but it never took off.

  4. #4
    notoriousjim's Avatar
    notoriousjim is offline
    Join Date
    Mar 2008
    Posts
    1,538
    Location
    maryland
    Rep Power
    19

    Dobber Sports Stud

    Default

    a few things that surprised me:
    centers are better at SOG that i expected. I guess i assume the prototype C is H Sedin. Great assist and power play numbers, but weaker in goals and SOG. I guess the prototype is further from sedin and moving closer to a guy who shoots as much as he sets up.

    Aside from Z score, you should also look at average stats, or stats over 82 games. Centers look weak with a short crosby season. Also i would go by last season rank. Skinner and a few others had down years, but are getting drafted based on upside rather than who they are.
    10 team Dyn
    tavares rnh statsny
    AO benn
    kessel malkin
    ryan purcell
    green letang enstrom kulkov OEL ellis
    lundy ward varly bernier lehrner
    burmistrov hodgson michalek, ryan oreilly, turris okposo

    16 team h2h
    Spezza, Tavares
    Vanek,doan
    malkin, eberle
    Doughty Letang, karlsson, weber
    hall
    green enstrom downie palmeiri kassain bogo
    lundy price pavelec

    weiters fielder kipnis wright hanley braun upton upton maybin victorino goldschmidt
    kershaw, startsburg, greinke, bumgarner, price, neise, bret anderson
    valverde, hanrahan, aceves, reed, guerra, storen

  5. #5
    Location
    Toronto
    Rep Power
    50

    Dobber Sports Star

    Default

    Just as a little plug, rather than have Ma come in to the rescue to dig up the data all the time (), this exact output is generated as part of the League Breakdown tool on FHG.

    The League Breakdown takes the population of owned players based on the league size, and generates tables summarizing both the average player at each position and the lowest owned (ie waiver replacement-level).

    For each player with multiple eligibility, they are included in the calculation for the position where they are more valuable.

    It doesn't give a player-by-player approach, but that's what our rankings are for -- values are based on standard score (z-score) in each category when compared against the replacement player at their position. Using this valuation approach systematically maximizes overall value on your team, though you still need to manage for your individual categories.

    End plug!

  6. #6
    Location
    Ontario
    Rep Power
    40

    Administrator

    Default

    Quote Originally Posted by fantasyhockeygeek View Post
    Just as a little plug, rather than have Ma come in to the rescue to dig up the data all the time (), this exact output is generated as part of the League Breakdown tool on FHG.

    The League Breakdown takes the population of owned players based on the league size, and generates tables summarizing both the average player at each position and the lowest owned (ie waiver replacement-level).

    For each player with multiple eligibility, they are included in the calculation for the position where they are more valuable.

    It doesn't give a player-by-player approach, but that's what our rankings are for -- values are based on standard score (z-score) in each category when compared against the replacement player at their position. Using this valuation approach systematically maximizes overall value on your team, though you still need to manage for your individual categories.

    End plug!
    I'll keep the plug going. This stuff is awesome. That is all.


    Contact me for Frozen Tools bug reports and inquiries
    Follow Frozen Tools on Twitter @FrozenTools
    Follow me on Twitter @DH_EricDaoust

  7. #7
    Dakkster's Avatar
    Dakkster is offline
    Join Date
    Aug 2007
    Posts
    11,669
    Location
    Halmstad, Sweden
    Rep Power
    50

    Dobber Sports Juggernaut

    Default

    Quote Originally Posted by Pengwin7 View Post
    Players should always be placed at the weakest position for analysis.
    For this year, I've judged the following:

    C/LW is a LW
    C/RW is a RW
    LW/RW is a RW
    RW/D is a D

    So... THIS year, RWs are as strong as ever.
    When/if the statistical providers ever get their positional act together, then guys like Malkin, Giroux, and Backes should all be C only. Guys like Zetterberg, Datsyuk, Benn, M.Richards, Ott should also be C only (not C/LW).

    Here is a link to all dual Yahoo! guys for this year 2012-2013.

    Also, when I was tracking player-for-player, I used multiple counters to sum the total players at any forward position and also sum them in my preferred position (per above).
    That thread is here... but it never took off.
    Small thing, but Ott is a leftwinger, not a center. He just takes faceoffs.
    GO WINGS!

  8. #8
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default

    I thought about using z-scores, but the thing with adding up of z-scores, the problem is if you have 1 very very negative stat that could completely change the numbers.

    Eg.) if you have +1 G, +1 A, +1 PPP, +1 SOG, but they have very terrible +/- and PIM numbers, they would be painted very negatively, which is kinda what FHG does, cause it punishes for negative scores. So even though Kessel is great for 4 out of 6 stats, but when you factor in the negative factor for +/- and PIM, he slides down to 76th... That's the downside with z-scores.
    Last edited by Maaaasquito; November 13, 2012 at 4:53 PM.
    Dobber Expert Pool Champion 2011-12

  9. #9
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default

    Quote Originally Posted by Skin Blues View Post
    You can also run a few iterations on the whole pool of players and find out who the best 14 are at each position rather than just using the top 14 drafted, which are almost always not the best 14 players.
    Yeah the only problem is because everyone's draft lists are different, so coming up with the top best 14 at each position, might not actually end up that way. What generally wins pools is getting players well below their draft value, but they end up producing top-tier numbers. Eg.) Landing Karlsson last year with a mid-round pick, was the difference maker in most leagues, or Perry the season prior.

    So I kinda just went with the Yahoo! ADP instead, generally speaking the top-14 drafted players at each position is relatively stable, it's just a matter of who ends up where?

    But you are right there are multiple ways of looking at things, I just went with the ADP route.
    Dobber Expert Pool Champion 2011-12

  10. #10
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default

    Quote Originally Posted by fantasyhockeygeek View Post
    Just as a little plug, rather than have Ma come in to the rescue to dig up the data all the time (), this exact output is generated as part of the League Breakdown tool on FHG.

    The League Breakdown takes the population of owned players based on the league size, and generates tables summarizing both the average player at each position and the lowest owned (ie waiver replacement-level).

    For each player with multiple eligibility, they are included in the calculation for the position where they are more valuable.

    It doesn't give a player-by-player approach, but that's what our rankings are for -- values are based on standard score (z-score) in each category when compared against the replacement player at their position. Using this valuation approach systematically maximizes overall value on your team, though you still need to manage for your individual categories.

    End plug!
    I could be wrong, but when I was playing around with the tool, I only got the average numbers per position, so all of the RW, or all C, but there wasn't a way to distinguish a #1 RW from a #3 RW... So using the average numbers of the RW position is ok, but the numbers I want from my #1 RW is different than the numbers I'd expect from my #3 RW.

    Is there a way to break it down per position? Like top 10-14 RW, then next 10-14 RW then bottom 10-14 RW?
    Dobber Expert Pool Champion 2011-12

  11. #11
    Location
    Toronto
    Rep Power
    50

    Dobber Sports Star

    Default

    If you go to the league breakdown, you'll see there are two tables, one for average and one for replacement.

    You'd have to pull that data from the rankings themselves, I guess.

    I'm not sure what valued you'd get from comparing against different RW slots (#1 RW, #3 RW, etc) though... either way you want a single "reference point" against which to compare a player, regardless of which RW spot they fill.

    ... That's the downside with z-scores.
    Hah, funny how perspective works -- I'd argue that's the advantage with z-scores. If a player is soooo bad at filling up a given bucket, even if they're filling a few other buckets, it should be a major hit to their value. Downplaying those "knocks" has a massive cumulative negative effect on the performance of a team.

  12. #12
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default

    Quote Originally Posted by fantasyhockeygeek View Post
    If you go to the league breakdown, you'll see there are two tables, one for average and one for replacement.

    You'd have to pull that data from the rankings themselves, I guess.

    I'm not sure what valued you'd get from comparing against different RW slots (#1 RW, #3 RW, etc) though... either way you want a single "reference point" against which to compare a player, regardless of which RW spot they fill.



    Hah, funny how perspective works -- I'd argue that's the advantage with z-scores. If a player is soooo bad at filling up a given bucket, even if they're filling a few other buckets, it should be a major hit to their value. Downplaying those "knocks" has a massive cumulative negative effect on the performance of a team.
    Yeah definitely... but with that said, most leagues aren't cumulative based, you don't get 1 overall z-score as a score for the week, you break it down category by category...

    So if I get 1 more goal than my opponent in H2H I win that category, so it's not about how much above the standard deviation I am... I can be 3 s.d. above in 3 categories, but 0.1 s.d. below in the other 3, and I still 50/50 my win/loss ratio, but my cumulative s.d. for a player would be huge because I was so high in the 3 but low in the 3 losses... or the other scenario where I might have a player that is really crappy in 2 stats (-3 s.d.), but good enough to secure me 4 stats (0.1 s.d), the overall picture would paint him to be a bad fantasy player, but in a H2H scenario he might be a "good own" cause he helps me in 4 categories at the cost of just 2...

    which is where a player like Sedin for instance could provide you with such great value because in the areas that he is so strong in, he's so far above the rest of the alternatives that you could single handedly secure a couple of stats as opposed to being "across the board".

    I find Thornton the same boat, which is why I'm constantly left scratching my head why his value he's constantly under-rated... Yes his G, SOG are very low compared to his tier of Cs but his A, +/- and PPP are so far above the average that it gives him such an edge that you're trading 3 stats for 2... In H2H it's about winning half + the tie breaker to secure victory, you don't need to win anymore than that.

    Roto is a little bit different, because you are after higher "highs", but you also don't want low "lows" as well. You have to have an "across the board" team. So you kinda want to have players with higher standard deviations... but there's also a limit on those standard deviations as well, like you don't want a team that crushes everyone in goals, but you are crap in assists, +/- and PIM...

    For me I want to secure enough to get me the victory, and grabbing the best player that will get it done for me.

    That's where I find z-scores a bit of a dilemma, because you're adding all the z-scores to paint an overall picture of a player, but in fantasy, player values are judged by individual categories as opposed to a cumulative score. A scenario like Kessel would be great for H2H cause he helps out in 4 categories while sacrificing 2, which would lead to a small victory. The only problem for me is the 4 areas in which Kessel is strong in there is leeway for error, whereas an option like Sedin, because the areas of his strength are so much more than the average of the alternative, I can lock his strengths away and not have to worry too much about leeway for error.
    Last edited by Maaaasquito; November 13, 2012 at 8:52 PM.
    Dobber Expert Pool Champion 2011-12

  13. #13
    blayze's Avatar
    blayze is online now
    Join Date
    Sep 2010
    Posts
    15,241
    Location
    Toronto
    Rep Power
    50

    Dobber Sports Ninja

    Default

    Great analysis Ryan, very interesting stuff.

    This confirms what I've known all along... that Nash SUCKS

  14. #14
    Maaaasquito's Avatar
    Maaaasquito is offline
    Join Date
    Dec 2007
    Posts
    3,449
    Location
    Melbourne, Austra
    Rep Power
    39

    Dobber Sports Expert

    Default

    I dunno Nash is an interesting case, I don't think anyone can argue if the talent is there, but he's also never really had much help in CLB either.

    Going to NY he'll have Richards and Gabby plus Callahan, Stepan, Kreider, Hagelin... so he'll have much more offensive support. With that said, he also isn't going to get the ice-time that he did with the Blue Jackets (19:05 and 3:27) in NY, so that might be a leveling factor as well.

    Keep in mind that Gabby only hit 76 points last season and Richards only hit 66, so I mean you have to keep Nash's numbers in check and not expecting something overly unachievable as well.

    One thing that's interesting with the Rangers is that under Torts they've been a very consistent team. 2.71 GF/G last season, 2.73 the year before, and 2.67 the season before that, now there's no evidence to prove that they have to be within that range heading into this season, but if you follow the trends that's probably where I would peg them. They did lose Anisimov's 36 points and Dubinsky's 34 points and Fedotenko's 20 points, so I mean there's 90 points to spread back around elsewhere within the team. Let's say Richards picks up a few more points (10-15), Nash posts 65-70 that essentially means everyone will stay status quo if the Rangers keep around the 2.70 GF/G mark. That's probably how I would approach it.

    If I were a Nash owner, I'd probably prefer that he stayed in CLB (and garner a ton of ice-time being the "go-to guy") and hopefully get a bit more help around him than for him to go to NY and become a "role player".
    Last edited by Maaaasquito; November 13, 2012 at 10:41 PM.
    Dobber Expert Pool Champion 2011-12

  15. #15
    Location
    Toronto
    Rep Power
    50

    Dobber Sports Star

    Default

    That's where I find z-scores a bit of a dilemma, because you're adding all the z-scores to paint an overall picture of a player, but in fantasy, player values are judged by individual categories as opposed to a cumulative score. A scenario like Kessel would be great for H2H cause he helps out in 4 categories while sacrificing 2, which would lead to a small victory. The only problem for me is the 4 areas in which Kessel is strong in there is leeway for error, whereas an option like Sedin, because the areas of his strength are so much more than the average of the alternative, I can lock his strengths away and not have to worry too much about leeway for error.
    I think you need to be a little clearer here and draw an important distinction: value to your team is judged by individual categories. I'm not by any means saying you don't need to manage specific categories because you absolutely do, but I am saying that holistic player valuation happens separately from that. If you're running away with the A and PPP categories in a matchup and lacking in the PIM that'll seal the win, does that mean you should punt Henrik or Thornton for Konopka because the value to your team of Konopka is that much higher? Of course not! It's a simplistic example, but it demonstrates that "value to my team" and "player value" need to be separate concepts.

    You're certainly right that each category is measured by the individual stats and that the margin of a win is a single point/shot/PPP/etc but those can still be expressed in terms of z-scores (ie standard deviations within that stat from a point of reference).

    The inherent risk to managing as you suggest -- not giving full weight to extremely poor performance in "punt" categories -- is that your team is considerably more vulnerable to cold streaks and injuries. As long as you can accept/manage that risk you might be better off, but I'm not convinced.

    My view has always been that you can maximize value on your team (independent of team needs) while managing to category needs (as you need to) - it lets you have strong trade value to address unforeseen needs and gives a better ability to manage variation.

    Great discussion.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •