Page 1 of 2 1 2 LastLast
Results 1 to 15 of 20

Thread: 2013-2014 Yahoo! Rank by Category Score

  1. #1
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default 2013-2014 Yahoo! Rank by Category Score

    For third straight season, I've exported Yahoo!s hockey stats to Excel to see if I could match their "Rank" algorithm.
    Last year I pegged it perfectly.
    This year, not so perfect.
    There are 860+ listed skaters and there are a few strange numerical differences at certain ranks that I couldn't fix.
    Bah.
    (Explanation: Each player's gray category score should be lower than the player above him in rank. Where there isn't the case, by my spreadsheet, there is a number showing the difference. I adjust pink cells manually, tediously, to try to minimize the sum of the differences. Last year I completely ironed it out to a perfect zero-error list... but this year, there are weird ones. Two noticeable ones occuring at #54 Kyle Turris value-too-high and #59 Clarke MacArthur value-too-low. ???)

    Anyways - for those that like the in-depth numbers, attached is my PDF of how Yahoo (might) compile the "Rank" based on category scores.
    Yahoo's base 6 categories are: G, A, +/-, PIM, PPP, SOG.
    Each player gets a weighted score in each category:

    For example (for league leaders of top 50 skaters, say):
    80.863pts --> Ovechkin's 51 goals
    90.218pts --> Crosby's 68 assists
    38.000pts --> Bergeron's +38 [I use this as my "BASE" for my "score". +1 = 1pt. -1 = -1pts. Everything is weighted -see pink cells- off this.]
    51.844pts --> Backes' 119 PIM. (not league leader, but you get the point)
    79.717pts --> Backstrom's 44 PPP.
    103.87pts --> Ovechkin's 386 SOG.


    Category scores are good to understand, because you can see what "increase" or "decrease" in value a player might have if a certain statistic changed.

    For example,
    Ovechkin was a -35.
    If he had simply been a +0, he'd have +35 more points.
    That would put him at 313.450pts, still short of Crosby, but it would make Ovechkin a consensus #2 player in fantasy hockey, which is where he'll still likely be drafted.

    Corey Perry was a +32.
    If levels out to something like +10, you could remove -22pts.
    That would send him down to 268.196pts... placing him below Benn/Seguin/Pavelski.


    ***This is for FULL-SEASON stats, exported straight out of Yahoo!***
    So obviously guys that were injured like Stamkos & Malkin are going to be much lower in "Rank" than they would for a full 82-game schedule.


    Enjoy.
    -P7
    Attached Files Attached Files

  2. #2
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    A follow-up to this spreadsheet, as there is a lot of good information in there.
    Brent Burns was D-eligible in Yahoo! this year.

    Even missing 13 games, he finished 7th among D in this multi-cat format.
    And he finished 4th in per-game production.

    There were a lot of good pre-season threads about Burns' value being D-eligible in Yahoo! leagues.
    He was pegged down on the Yahoo "O-Rank" (which is where they list guys in the "draft list") at #167... but finished at #62 overall.

    So BIG value for anybody that joined in these discussions and was able to draft Burns between about #80-#150 in their drafts.

  3. #3
    austeane's Avatar
    austeane is offline
    Join Date
    Sep 2010
    Posts
    1,495
    Rep Power
    22

    Dobber Sports Ace

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Awesome analysis! It's interesting looking at how they do their rank, versus how I do my top 100 roto.

    They do theirs based on weighting the categories differently, while I do mine with equally weighted categories and separating the players based on how they do in each cat compared to other fantasy-relevant skaters.
    Data Science Intern, Pebble
    COO, Quidditch Post
    Associate Editor, Dobber Sports Network
    Hockey Operations Intern, UBC Thunderbirds

    Top 100 Roto
    Dobber Prospects

    Ask me your fantasy questions on Twitter: @austeane
    My main league:

     

    H2H Each Cat - Yahoo - 14 teams - Real Cap Hit - 85m cap
    Start: 2C 2LW 2RW 4D 2U 2G 7B daily (21)
    Cats: G A +- PIM SOG PPG PPA SHP GWG FW HIT W
    GAA SV SV% SO
    Edler's Mind Tricks:
    C:, Lehtera, Plekanec, Kuznetsov, Komarov
    LW: Benn, Kreider, Roussel
    RW: Malkin, Yakupov, Nyquist
    D: Subban, Letang, Karlsson, Jones, Dillon, Deslauriers
    G: Bobrovsky, Lehner, Elliott
    JV: Sekac, Petan, Poirier, Vrana, Fleury, Sanheim, Galiev, 5th, 9th, 15th, 18th, 24th overall amateur picks

  4. #4
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Quote Originally Posted by austeane View Post
    Awesome analysis! It's interesting looking at how they do their rank, versus how I do my top 100 roto.

    They do theirs based on weighting the categories differently, while I do mine with equally weighted categories and separating the players based on how they do in each cat compared to other fantasy-relevant skaters.

    It is interesting.
    Because category weighting is completely subjective to the league.

    Each league really needs a specific weighting depending on how many players are "counting".

    I actually have some cells near the top right with a percentage.
    In those cells, I total up the percentage of stats that the top 144 skaters make up from the entire NHL.
    (Most common skater quantity is about 144. 12x12 (8F/4D approx). But different league sizes would make a difference in category score.)

    The most interesting is PPP.
    60% of total NHL PPP can be found in the top 144 skaters.

    Thus, in deeper leagues - teams should explicitly be going after PPP... because they "dry up" fastest.

    Category weighting can also skew based on how a league of GMs drafts.
    If GMs are NOT drafting PIM players, then a GM could get HUGE value from drafting one big PIM player to stock his PIM cat.

    i.e. If 12 teams draft an average of 100 PIM each, then drafting 120 PIM is a smart move.
    However, if all 12 teams are drafting 200 PIM each, than the total PIM in the league is MORE making the PIM-category-value for a certain amount of PIM worth less.


    A category IS just a category and (I agree) they should be weighted equally.
    So the key, for any league, is to determine the amount of statistics within the number of counting players.
    Then the weighting goes directly with the total statistics.
    Fewer total stats = more weight per stat.

    In the Yahoo! above, you can directly see my weights:
    One G = 1.59pts
    One A = 1.33pts
    +1 = 1.00pts
    1 PIM = 0.44pts
    1 PPP = 1.81pts
    1 SOG = 0.26pts

    In shallow pools, say a 10-team x 6F/4D pool, all of the drafted players will get PPP though.
    It becomes more well stocked than in a deeper pool and thus the PPP value per unit (per stat) would be less.

    A great number-crunching topic though.
    Weighting per Category vs. LeagueSize/RosterAllocation

  5. #5
    jross's Avatar
    jross is offline
    Join Date
    Jan 2013
    Posts
    168
    Rep Power
    0

    Dobber Sports Prodigy

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Just wanted to say Pengwin you're an excellent poster on these boards. You always have solid insights backed up by detailed posts. It's a pleasure reading your posts.

    I exclusively play fantasy-point-only leagues, so they category analysis is somewhat lost on me, but still interesting. Each of the five $$$ leagues I play have varied fantasy-point values for a variety of category's. It's a lot to keep track of.
    11 Team Keeper - Yahoo Pts Only
    G 4 - A 3 - +/- .5 - HIT .2 - SOG .2 - PIM .2 - BLK.2
    W 5 - L -3 - GA -.5 - SV .2 - SO 5

    C: Crosby, Seguin, Koivu, Dubinsky
    LW: Neal, Benn, Pacioretty, Dupuis
    RW: Eberle, Franzen, Vrbata, Atkinson
    D: OEL, Goligoski, Wisniewsky, Dillon, Faulk, Kulikov, Souray
    G: Backstrom, Lehtonen, Price, Khubodin
    Util: Grabo, Tlusty, Ruutu
    BN: Killorn, Clutterbuck, Bouchard, Schwartz

  6. #6
    Eyenerd's Avatar
    Eyenerd is offline
    Join Date
    Jan 2013
    Posts
    401
    Location
    Washington, DC
    Rep Power
    13

    Dobber Sports Blue-Chipper

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Awesome spreadsheet Pengwin7! Is there any way you can post it as an excel format so we can play with the weighting and sort the list according to the cats we need for our leagues? ie. my league typically doesnt count FOW or Blocks.

    I should start creating a draft list for next season
    1 Year H2H (G,A,+/-,PPP,GWG,PIM,Hits, SOG,W,GAA, SV%,SHO), 8 Teams 3C, 3LW, 3RW, 3F, 6D, 3G, 4 Bench, 2 IR

    C: M.Koivu, Sharp (LW), Stepan
    LW: Perron (RW), Soderberg, Landeskog
    RW: Ovechkin, Carter (C/LW), Skinner (C/LW)
    F: Simmonds, Johansen, Anisimov
    D: Doughty, OEL, Franson, Gudas, Enstrom, Daley
    G: Price, Varlamov
    BN: Stempniak, Perreault, Lee, Michalek, Maroon
    IR: Letang
    Lost to Injury: Tavares, M.Smith, Dupuis, Fowler, Parenteau, Kreider, Granlund, Lupul, McDonagh

  7. #7
    Top_Shelf's Avatar
    Top_Shelf is offline
    Join Date
    Mar 2013
    Posts
    357
    Rep Power
    16

    Dobber Sports Blue-Chipper

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Can't believe I just saw this now...awesome work Pen!

    +1.....I`ll send you rep when I can again.

  8. #8
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Thanks guys.
    Yes - the spreadsheet can also be tweaked to include other categories like FOW, HIT, BLK.
    I've had a bit of trouble pegging FOW value, because a LW or RW should have an average FOW that is very low.
    And with Yahoo!s ridic pos.elig., slotting a Giroux or Backes in a wing spot gives them an incredible spike in one category... but that all depends on how other teams in a league deal with FOW.

    Overall, I find the spreadsheet most valuable for identifying players that may have had strangely low scores in a correctable category.
    For example, Ovechkin at -35 is something like to correct a bit.
    And guys with low PPP totals (say, Jamie Benn) are going to be worth even more if they can put up 20-25 PPP.


    Even more interesting than "skater score" is "goalie score"... which I still struggle with.
    But maybe this year we have a few extra smarties on the board to exchange ideas with about how to accurately put a category score on the range of GAA & SV% (very, very, very difficult topic).

    Thanks for comments!

  9. #9
    Location
    Hamilton, ON
    Rep Power
    40

    Dobber Sports Titan

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    With goalies it looks like past years are incorporated. Rinne in 2013 didn’t have a huge yr, but is ranked higher than Niemi and Anderson (who was great in 2013). And wins don’t have as big as an effect on the ranking. No1 ranked goalie was Rask with 19W, No2 was Lunquist with 24W, No3 was Rinne with 13W, No4 was Neimi with 24W, and No5 was Anderson with 12W.

    All this is a Lockout Year, so maybe they adjusted their math for rankings and went more of a freelance system based on reputation. Like Rask total points x 1.35 based on youth, experience, and lack of injury history, Lunquist total points x 1.25 based on constancy and reliability. I’m sure games starting is in there too, Rinne playing 70 games makes it a higher chance for him to put out 30+ wins, more saves, more chances for shutouts.

    BTW, awesome chart. I will be using that as a point of reference for sure
    12 Team, H2H, Keep 6 (in Bold)
    G, A, Pts, PPP, FW, SOG, Hits, Blocks
    W, Saves, S%, GAA, Game Started
    2C, 2LW, 2RW, 4D, 1Util, 2G, 5BN, 2IR, 1IR+, 1NA

    C: Horvat, Trocheck
    LW: J. Robertson, Byfield (C), Guenther
    RW: Pavelski (C), Giroux (C), Svechnikov (LW)
    D: Fox, Makar, Bouchard, Morrissey, Gudas
    Util: Meier (LW, RW)
    G: Oettinger, Georgiev, Samsonov, Woll


  10. #10
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    ^I've always been annoyed with Yahoo! because they definitely do some sort of projections/predictions to set up their start of season O-Rank (their base list for drafts).

    A few years ago, I remember they had it weighted purely by past stats... and Crosby showed up really low, along with other past-injured guys.
    Since then, they've come into the season with pretty good placement for players, but they've still missed a few. Dubinsky was one this year that smart managers feasted on, he was O-Ranked in the 200s or 300s I believe, because he only played 29 games in the 48 game lockout season... and I think Yahoo! improperly took that without extrapolating.

  11. #11
    Location
    Hamilton, ON
    Rep Power
    40

    Dobber Sports Titan

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Varly was projected really low for this season too, especially for a starting goalie. I think Bernier was actually ranked higher than Varly. There's something different they do for goalies. Maybe just a panel of people saying one goalie is better than the other. The top 10 goalies even seem a little oddly done when just looking at the previous yrs stats. Rinne over Neimi is a bigger kicker for me.Rinne's history to me would of put him in front, but not based on last yrs stats.
    12 Team, H2H, Keep 6 (in Bold)
    G, A, Pts, PPP, FW, SOG, Hits, Blocks
    W, Saves, S%, GAA, Game Started
    2C, 2LW, 2RW, 4D, 1Util, 2G, 5BN, 2IR, 1IR+, 1NA

    C: Horvat, Trocheck
    LW: J. Robertson, Byfield (C), Guenther
    RW: Pavelski (C), Giroux (C), Svechnikov (LW)
    D: Fox, Makar, Bouchard, Morrissey, Gudas
    Util: Meier (LW, RW)
    G: Oettinger, Georgiev, Samsonov, Woll


  12. #12
    Location
    Hamilton, ON
    Rep Power
    40

    Dobber Sports Titan

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Book1.pdf OK bear with me on this one. I did a bit of monkey-ing around with FW, and pro-rating injured player’s numbers. I also did a position bonus, and winger face-off bonus. So Giroux, who has 2 positions, got a bit of a boost, he also got a bonus because he was a “winger” who got FW. So Giroux got 2 bonuses. However, a player like Taylor Hall has 2 positions (C/L), but doesn’t take face-offs only got 1 bonus. Paves, (C/L/R), got 2 bonuses for positions, and 1 bonus for “winger” face-offs for a total of 3. Keith and Karlsson also got a bonus for being D. Stamkos, Tavares, Malkin, and Neal all got their numbers fudged a bit. I took their average ranking points per game (A Player was let’s say 150pts over 60 games; I took that and made it to a realistic games played number of 70. So (150/60)*70realistic games=175). Stamkos and Tavares were pro-rated to more games than Malkin and Neal because of their history. So this chart takes into account:
    - Goals
    - Assists
    - PPP
    - +/-
    - FW
    - PIM
    - Positional Bonus
    - Winger FW Bonus
    It’s super crude, but you can see how Giroux jumps from 3rd to 2nd, Getzlaf from 2nd to 9th, Paves from 11th to 3rd, Seguin from 4th to 17th, Perry drops a lot too. I think I have to play with the numbers a bit considering Hall is below Simmonds somehow….and Ovi from 8th to 27th, or how Karlsson is now below Wheeler. I might of given FW too much of a bonus.

    But it’s cool to see whose worth more than whom. Or how only 2 people in the top 23 have a negative +/-. Or how on the new list I made the top real winger that played the season is Perry at 16th.
    12 Team, H2H, Keep 6 (in Bold)
    G, A, Pts, PPP, FW, SOG, Hits, Blocks
    W, Saves, S%, GAA, Game Started
    2C, 2LW, 2RW, 4D, 1Util, 2G, 5BN, 2IR, 1IR+, 1NA

    C: Horvat, Trocheck
    LW: J. Robertson, Byfield (C), Guenther
    RW: Pavelski (C), Giroux (C), Svechnikov (LW)
    D: Fox, Makar, Bouchard, Morrissey, Gudas
    Util: Meier (LW, RW)
    G: Oettinger, Georgiev, Samsonov, Woll


  13. #13
    Location
    Prairies
    Rep Power
    40

    Dobber Sports Demi-God

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    ^That's the way my numbers shook out last year as well.
    Centers got a FOW category score based on what they provided relative to the average center.
    Wingers got a FOW category score based on what they provided relative to the average winger.

    And yes, last year, it pushed my wing-eligible centers WAY up the rankings.

    Two schools of thought on this:
    1. In Roto, the value is there, but not beyond what is needed to get you the 1st place total. (In my Roto league this year, I believe the top three teams finished top three in FOW, using wing-eligible centers).
    2. In H2H, the value is there, but it is very limited, considering it gives you a "Vs. 1 team advantage"... and not a "Vs. all team advantage".

    Good numbers, looks like I have found a numbers-buddy to throw stuff back and forth with.
    Next week - we attack goalie scores!!!

  14. #14
    Location
    Hamilton, ON
    Rep Power
    40

    Dobber Sports Titan

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    oh god this is going to be scary^. better tell my boss i'm going to be mentally checked out
    12 Team, H2H, Keep 6 (in Bold)
    G, A, Pts, PPP, FW, SOG, Hits, Blocks
    W, Saves, S%, GAA, Game Started
    2C, 2LW, 2RW, 4D, 1Util, 2G, 5BN, 2IR, 1IR+, 1NA

    C: Horvat, Trocheck
    LW: J. Robertson, Byfield (C), Guenther
    RW: Pavelski (C), Giroux (C), Svechnikov (LW)
    D: Fox, Makar, Bouchard, Morrissey, Gudas
    Util: Meier (LW, RW)
    G: Oettinger, Georgiev, Samsonov, Woll


  15. #15
    Location
    Avalanche Country
    Rep Power
    13

    Dobber Sports Prodigy

    Default Re: 2013-2014 Yahoo! Rank by Category Score

    Pengwin you a stud. Can you post in Excel as well? It would be great for each user to manipulate the inputs based on individual league settings.
    12 Team H2H. Full Keeper
    G/A/Points/PIM/PPP/FOW/SOG/HIT/BLK/W/GAA/SV/SV%/SO
    Malkin, Seguin, Perry, Monahan, TJohnson, Wheeler, Lucic, Nichushkin, Drouin, Datsyuk, Hossa, Holland, Bouma, Nieto, Colborne, Prust, Jooris
    Subban, Savard, Dillon, Murphy, Prout, Petrovic
    Lundqvist, Rinne

Posting Permissions

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