News:

(**folks need to register to be able to see Fibsboard Forum Match diagrams and vote.** )

Main Menu

TL4 Pre-season Odds

Started by sixty_something, October 09, 2007, 07:05:45 PM

Previous topic - Next topic

sixty_something

before play started on Team League for this season, i pulled the FIBS ratings for each player .. from this i calculated a simple average rating for each team .. using this average team rating, i then developed a simple stochastic simulation of season matches on a team by team basis in Excel .. using the average rating for each team, it is then possible to calculate the probablity the underdog might win the match

while the team approach does not account for differences by position, it is an interesting approximation and insight into the race for the new season championship .. here are the odds for each team to finish first based on a simulation run for 10000 seasons:

Quote
Odds   Team                              Avg Rating
   7:1   Smack, Slap, and Spank      1849.53
   8:1   Club Voodoo                          1840.86
   8:1   Cube Sluts                             1833.09
   9:1   Ottoman's Empire                1825.19
11:1   TakeTheMoneyAndRun        1804.88
13:1   Lipstick & Testosterone      1789.12
13:1   Virtual Addicts                      1785.74
13:1   Gammonizers                        1783.27
17:1   No Nuts A--hole                    1761.90
19:1   Last Minute                           1752.89
23:1   schigstogsedat                    1737.84
31:1   TBD                                        1714.03
44:1   MOJO                                     1649.88
47:1   UN Trio                                   1672.54
51:1   three.DOG.night                  1677.19
65:1   white punks on dope           1624.33
239:1   two dicks and a chick         1525.60

we've got four strong teams battling at the top and a strong field .. anything can happen, fun most assuredly will .. an interesting observation was that over 50% of the simulations resulted in a tie for first place between two or more teams .. yet another indicator of a close race .. good luck to all and enjoy .. thanks to souptree, zyxtcba, and all the rest who've made this possible

sixty
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me

Tom

So are we doing side bets?? :)

Will you pay out on those odds?

tom

Sequit

Hi 60.

If you did your ranking within the steps, and then averaged each team's ranking, how would that change your odds? I'd do it for you, but this post took my free time for this week.

sixty_something

#3
the parimutuel window just closed, Tom .. heavy last minute betting on two dicks and a chick brought their odds down from 750:1 (in a 3000 season simulation) to 239:1 (in 10000 seasons) .. so, they could be a good show bet .. anyway, the race is on with a new team now, Last Orders with blotsalot, padski, and erdbeere .. it looks like they'll go off at about 55:1

so, maybe next season for parimutuel betting in FIBS Yen using a funny money based on RepBot points, experience, and Tourney points .. why not, eh .. can you have that ready by Sunday?

Sequit, i expect a position by position simulation would change the odds somewhat, maybe not by much .. while each position plays out very differently, gains in one position bracket would tend to be offset by losses in another .. differences of 100 rating points change the probability calculation of the underdog winning by about 0.100 or 10% .. since the ratings of individual players by position vary much more than team average and individual wins contribute to team points, positional simulation would be a much more accurate approach, but whether it makes a big difference we'll just have to wait and see

my seat-of-the-pants estimation tells me it won't make that big a difference in preseason odds or even how the order comes out.. how the dice fall and the matches play out will do that big time as the season gets under way .. regardless, i'll try doing this again at the quarter poles, halfway mark and down the stretch updated with actual results at each checkpoint .. that may prove interesting

i tried running a positional simulation this morning, but the way i setup the team by team simulation in Excel requires more tinkering, time, and talent to get it done than i had this morning .. maybe by the update at the quarter pole

boomslang has been working on a statistical approach which may be an easier way to go there anyway .. others are suggesting we update the data continuously .. pretty soon we won't need to play at all, we'll just sit around simulating ourselves, but personal experience tells me it's not nearly as stimulating or satisfying  :unhappy:

sixty
A little inaccuracy sometimes saves tons of explanation. -- Unknown
e-mail me