Forum
Dottir takes November TAZD.
|
skrumgaer wrote
at 9:57 AM, Thursday December 1, 2011 EST
The TAZD and baseball-style standings are explained on my Wall. At least 35 regular games played in the month are require to qualify for the monthly TAZD. Shown are Games Behind, TAZD, and player name.
GB TAZD Player 06 12178 dottir 13 11141 Emre Oguz 03 10171 masticore 00 9719 Invola 39 9539 Shevar 03 8878 OneShot7 18 8842 jona_vicente 06 8419 savif 22 8352 [Ocean]Flushed 32 8336 Mazaman 02 8224 toms 10 8170 what_up23 47 8155 jfdis 08 8113 @ata 24 8064 Az_Balu 17 7666 kostur 20 7604 L3xy 48 7603 bcmatteagles 16 7600 22-Apr 11 7427 Lady Lite 07 7406 Vollhonk 66 7294 Scabbard 26 7159 kdiceplaya! 22 6840 chaiNblade 29 6829 IFIGENIUS 17 6518 FPP 24 6504 _smile_ 69 6474 Remiel 43 6441 Simmo3k 40 6411 Mercantile 12 6397 xjxaxnx 11 6328 @Toomyfriends 93 6315 franklyghost 14 6259 Bu7Ch3r 34 6214 fish28 18 6129 Free Flags 19 6043 hcdug 24 5928 kudoukun 18 5921 ovbogaert 14 5907 peter luftig 36 5658 @engr2002 49 5588 EddyB 22 5474 @MikeTamburini 31 5398 Brighty 30 5333 fearlessflyer 39 5281 Lord Death 92 5210 Loobee 35 5123 Gurgi 66 5087 barmat 21 5065 joero14 66 5054 Jily 40 5044 hatty 33 4952 longpube 32 4921 NikkeKnatterton 29 4841 scarp8 54 4794 stackshotbilly 34 4784 OviloN 66 4733 Silesia 100 4730 axlehammer 45 4623 mrb2097 47 4600 nexon 21 4582 Volvic 23 4484 beatol 33 4471 Fatman_x 25 4411 KDancer 41 4306 xXxJozefxXx 25 4289 Keeley 26 4019 euphrates7 87 4003 Rsquared 36 3917 Poker Style 48 3808 "MC" 34 3760 haloducks 41 3641 bivo 69 3261 orestis85 52 3201 greekboi 73 3179 cool g 33 2960 MNK10 57 2817 Trkz 58 2784 greenman 65 2759 These tards suck 76 2714 GreGGwar 70 2500 absolutgimlet 61 2463 Johnboat 44 2285 Kingofskillz 84 2218 DonnieScribbles 93 2208 GR3ENMAN 73 2028 CCSKAOT 94 1253 Kdot 92 1248 ji-jo |
|
|
skrumgaer wrote
at 12:46 AM, Sunday December 4, 2011 EST The number of samples is the number of games, ranging from as few as 720 to close to 6000 for the yearly TAZD.
On my wall, the discussion about the square roots is actually in chapter i, the TAPL chapter, although there is some corrupt text that makes it hard to read. |
|
|
skrumgaer wrote
at 12:50 AM, Sunday December 4, 2011 EST If sample size is not needed if you are using percentages instead of incidences, what do you think of my percentages for September 2008?
|
|
montecarlo wrote
at 12:51 AM, Sunday December 4, 2011 EST you are NOT comparing the number of games finished in each place. you are comparing the PERCENTAGES of finishes.
|
|
|
skrumgaer wrote
at 12:52 AM, Sunday December 4, 2011 EST I mean October 2008, it's late at night and I am getting bleary-eyed.
|
|
|
montecarlo wrote
at 12:54 AM, Sunday December 4, 2011 EST you only played 49 games, so your sample means have not overcome the uncertainty to resemble your true means.
or maybe you tried a different style of play? but the more games you play, the closer they will get to the true means, wherever the truth lied 3 years ago for you. |
|
|
montecarlo wrote
at 12:55 AM, Sunday December 4, 2011 EST wait, october 2008, as when you played only 1 game? we already went over this, you must have a statistically significant number of games before you can use the percentages to test with.
|
|
|
skrumgaer wrote
at 12:58 AM, Sunday December 4, 2011 EST You are comparing the square of the observed and expected number of finishes divided by the expected number. That is of the order of the expected number of finishes, that goes as the number of games. If you factor out the number of games in numerator and denominator, you have the ***number of games*** times the square of the difference of the expected and observed percentage, divided by the expected percentage.
Get some sleep. |
|
|
montecarlo wrote
at 1:12 AM, Sunday December 4, 2011 EST no, youre comparing percentages which are dimensionless. number of 7ths divided by total number of games. numerator and denominator cancel out. percentages are by definition dimensionless.
ill get some sleep eventually, but i cant sleep on this debate. |
|
|
skrumgaer wrote
at 11:27 AM, Sunday December 4, 2011 EST I am not comparing percentages, I am comparing incidences. Since Ryan gives the data in form of percentages, the percentages have to be muliplied by the number of games to get the number of incidences. Since the pecentage differences are squared in the numerators but not in the denominators, when you muliply the percentages by the number of games you get the square of the number of games in the numerators and the number of games in the denominators so the total expression goes as the number of games. Suppose you have 12 percent firsts and the expected value is 10 percent. If you have played 50 games, that would be (6 minus 5) squared divided by 5 or 0.2. If you have played 100 games that would be (12 minus 10) squared divided by 10 or 0.4. So the chi square goes as the number of games. Its variance also goes as the number of games, as I have said before, so its standard deviation goes as the square root of the number of games.
|
|
CCSKAOT wrote
at 12:37 PM, Sunday December 4, 2011 EST 150
|
