Jump to content


  • Content Count

  • Joined

  • Last visited

  • Days Won


Posts posted by andegre

  1. 56 minutes ago, nhs67 said:

    Did you happen to look at some of the squads on that league I did an autodraft for the other day?

    Some teams were absolutely jacked... others were wildly incompetent, yet I assume all 10(mine included in the autodraft) were done using the same sort of AI.

    Yes, when auto-pick is turned on, it will take the TOP wrestler, according to that user's "need weights", from the "Top 300/Best Available" list.


    If you check the "Top 300" page, that's the exact order of how they'd get drafted if auto-pick is turned on for everyone.


    But, the "Simulate Remainder of Draft", I believe, will choose, AT RANDOM, between the top 5 in the Need Weights/Best Available list...that was done to try and introduce SOME randomness when I test with all of the simulated users.

  2. 1 hour ago, Housebuye said:

    4. Renteria didn’t wrestle a match last year. Thanks thought he was likely to AA at 133 last year assuming he could make it, and that weight was stacked. Not sure what’s going on but it wasn’t an injury and he had classroom issues in the past. 


    It's been made public that JRent's father was going through chemo last season/year...that's why he wasn't wrestling.

  3. 6 minutes ago, buckeyehomer said:

    How does scoring work for this? Is it just a draft and you see how the season goes, and basically whoever gets the most points at ncaa's wins?

    It's not set-in-stone yet, but it will most-likely use dual scoring format, and it will be weekly head-to-head "matches" vs one other user in your league. Then the following week, it will be against a different user in your group. Winners will be determined by the W/L record, with tiebreaker being total points scored throughout the season.

  4. 1 hour ago, nhs67 said:

    Never mind. I believe I've found the issue. It won't allow any roster changes until the week before matches start 

    I don't believe that's the case.

    You ALWAYS [or at least should have] the ability to edit the roster for the following fantasy week, whether that's via starter changes, trades, or free agency transactions.


    For the trades, have you selected which other user in your league that you want to trade with? You have to do that, before you can see the rosters....



  5. 23 minutes ago, dakotajudo said:

    Suppose we assume that each weight class is an independent measure of ranking accuracy. How well do the different rankers predict the outcome of a bracket? If we assume that scores are continuous and that errors are randomly distributed, when we can test this using a simple AOV, in R (given that stacked.ching is the long version of the table Ching posted):

    ching.lm <- lm(Score ~ Source + Weight, data=ching.stacked)
    ## Analysis of Variance Table
    ## Response: Score
    ##           Df Sum Sq Mean Sq F value Pr(>F)    
    ## Source     8  187.0   23.38  0.7211 0.6722    
    ## Weight     9 9227.1 1025.23 31.6256 <2e-16 ***
    ## Residuals 72 2334.1   32.42                   
    ## ---
    ## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    Given that the data sums of scores, we might use a Poisson model,

    ching.glm <- glm(Score ~ Source + Weight, family=poisson, data=ching.stacked)
    ## Analysis of Deviance Table
    ## Model: poisson, link: log
    ## Response: Score
    ## Terms added sequentially (first to last)
    ##        Df Deviance Resid. Df Resid. Dev Pr(>Chi)    
    ## NULL                      89    218.387             
    ## Source  8    3.283        81    215.104   0.9153    
    ## Weight  9  173.040        72     42.064   <2e-16 ***
    ## ---
    ## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    We might continue with a categorical analysis of the table, perhaps Fisher's Exact test (here, the data are in original table form)

    fisher.test(ching.table,simulate.p.value = TRUE)
    ##  Fisher's Exact Test for Count Data with simulated p-value (based
    ##  on 2000 replicates)
    ## data:  ching.table
    ## p-value = 0.9975
    ## alternative hypothesis: two.sided

    Finally, if we don't assume any distribution, a rank-based test is appropriate

    friedman.test(Score ~ Source | Weight, data=ching.stacked)
    ##  Friedman rank sum test
    ## data:  Score and Source and Weight
    ## Friedman chi-squared = 3.4866, df = 8, p-value = 0.9002
    friedman.test(Score ~ Weight | Source, data=ching.stacked)
    ##  Friedman rank sum test
    ## data:  Score and Weight and Source
    ## Friedman chi-squared = 56.718, df = 9, p-value = 5.722e-09

    Long and short - there is more variation among scores between weight classes than there is variation among the rankers, and the range of differences among rankers is small relative to the error within weight classes. The difference between 544 and 588 is about 4 points per weight class, while high and low totals in any weight class, say 125 (45-65) varies more than that.

    There may be ways to further decompose the error in the comparisons to distinguish (at least in a statistical sense) among the rankers, but there's not enough information in Ching's table.


    Waymit! Did my "Oregon State alum who doesn't do social media" just get an account on this forum? lol

  6. On 3/24/2019 at 8:03 PM, Ching said:

    If any rankers want access to my spreadsheet with breakdowns, DM your e-mail address. I don't want to share publicly because the rankings are copyrighted.

    Hey @Ching, can you send me a couple of your excel file? My "Oregon State Alum who doesn't do social media" is already sending me modifications to start running simulations against....we're gonna get first eventually!!!

  7. It’s BASED/ORIGINATED from the Elo algorithm, with many changes to work for wrestling. There are approximately 10 different equations/algorithms that go into the whole thing. Then I wrote a simulation engine last off-season to allow me to do thousands of simulations to continually optimize it.

  8. Thanks again Ching!

    One thing that we/WrestleStat would also like to point out, if you exclude 285, WrestleStat was #1. (I know, picking data to fulfill my argument....).

    Pretty cool to see though, that all of the work myself and my algorithm guy did this offseason paid off in a BIG way. (Last year we scored around 515 if I recall).


    Thanks algorithm guy ;)


    Scratch that, leaving out 285 OR 157 and WrestleStat wins... :)


    edit 2: my “algorithm guy” wants his name changed to: “Oregon State Alum who doesn't do social media”, so thank you, “Oregon State Alum who doesn't do social media”

  9. 3 days to go!
    If you like using the WrestleStat comparisons, we've created a "Hub" page that will be used throughout all of Nationals this year.
    For each session, we'll be putting out these comparison pages, like the one linked above. Each time a new one gets released, we'll be updating this Hub page with the appropriate link.
    This way, you guys/gals can just bookmark a single page, and get to all the comparisons that we come up with. Should make it real easy while there so you don't have to hunt through message board posts...
    Here it is:

  10. 19 hours ago, SetonHallPirate said:

    I'm actually affected by it at 141, 149, 157, 184, and 285, for what it's worth. I wonder what weights the human rankers are affected.

    And here's his response:



         To answer SHP's question, for some of the human rankers, InterMat is affected 6 times, The Open Mat 6 times and Flo 7 times. For the computers, SHP is affected 5 times and Wrestlestat is still hurt the worst at 9 times.

                Intermat at         133,157,184 twice and 285 twice = 6 Total

                The Open Mat at    133,149,157,184,197 and 285 = 6 Total

                Flo at          125, 133, 141, 149, 157, 184 and 285 = 7 Total


    FYI, going back over a number of past Nat'l tournaments and averaging the data, here is where the 8 placers come from:  From the quarter finals, 4 winners place.

              From the quarter finals, 2 of the 4 losers place ( 4 x 0.5 = 2 since losing Q must win 16 - Q match to place).

              From  Rd. of 16  losers, 1 places ( 8 x 0.5 cubed = 8 x 0.125 = 1 since must win 32-16, 16-16 and 16-Q).

              From  Rd. of 32  losers, 1 places (16 x 0.5 to the 4th power = 16 x 0.0625 = 1, 32-32, 32-16, 16-16 and 16-Q).

    That last section shows the significance of this....quarters 2 of 4 place, rd16 losers only 1, and rd32 only 1.

  11. Doesn't look good for WrestleStat already this year, not just because of the rankings, but how the wrestlers are SEEDED... Check out what my algorithm guy sent me explaining why we'll do poorly this year...


    Hey Greg,

       We are in big trouble in the RTR contest.  9 times in the 10 weights, 2 of our top 8 guys meet each other in the round of 16 or 32 whereas this happens only one time for SHP.  Weight 184 looks bad also. This is serious because the chance of placing for the loser of the match is only 12.5% = 0.5 x 0.5 x 0.5 = 3 wins (16-32, 16-16  16-Q ) to place in the top 8 if you lose in the Rd. of 16 and 6.25% chance to place if you lose in the Rd. of 32. See below for details:


         133:   Our 5 and 8, seeded 7 and 10, meet in Rd. of 16   (7+10=17)

         141:   Our 7 and 5, seeded 9 and 24, meet in Rd. of 32   (9+24=33) 

                   Our 4 and 6, seeded 3 and 14, meet in Rd. of 16   (3+14=17)

         149:   Our 6 and 8, seeded 7 and 10, meet in Rd. of 16   (7+10=17)

         157:   Our 3 and 7, seeded 3 and 14, meet in Rd. of 16   (3+14=17)

         165:   Our 6 and 8, seeded 6 and 11, meet in Rd. of 16   (6+11=17) 

         174:   Our 1 and 7, seeded 1 and 17, meet in Rd. of 16    if our # 7 seeded 17 beats 16 seed in Rd. 32 (16 +17=33)

         197:   Our 7 and 5, seeded 8 and 9  , meet in Rd. of 16    (8+9 =17)

         285:   Our 5 and 8, seeded 6 and 11, meet in Rd. of 16   (6 +11=17)


         184:   Our 5 meets 9th seed in Rd. of 16

                   Our 6 meets 6th seed in Rd. of 16

                   Our 7 meets 4th seed in Rd. of 16

                   Our 8 meets our 10     in Rd. of 16


    He's sent subsequent emails clarifying that @SetonHallPirate is affected by this same scenario at 4 different weights (149, 184, 197, 285). Will be interesting now to see how the rankings hold up, though I'm not as confident as I was previously.

  • Create New...