Does anyone know how to solve for lambda in the Poisson equation below?
P(X=k)=
As of now, I'm using Excel's solver function to find the mean (lambda) given the market's odds for a Poisson distributed total.
Thanks
|
Full Poll Results
|
SBR Book & Casino
|
|||||
| #1 5Dimes (SBR rating A+) | 294 Pts | #1 Tatddy | 14,300 Pts | SBR Sportsbook | ||
| #2 Pinnacle (SBR rating A+) | 264 Pts | #2 Studmlb55 | 10,522 Pts | SBR Sportsbook | ||
| #3 Bookmaker / BetDSI (A+) | 236 Pts | #3 Gators0708 | 5,100 Pts | SBR Sportsbook | ||
| #4 Heritage (SBR rating A) | 203 Pts | #4 Riskyprops | 4,600 Pts | SBR Sportsbook | ||
| #5 Bet365 (SBR rating A) | 78 Pts | #5 The Kraken | 4,300 Pts | SBR Casino | ||
Does anyone know how to solve for lambda in the Poisson equation below?
P(X=k)=
As of now, I'm using Excel's solver function to find the mean (lambda) given the market's odds for a Poisson distributed total.
Thanks
3-QUESTION
SBR TRIVIA WINNER 05/13/2013
3-QUESTION
SBR TRIVIA WINNER 05/16/2013
3-QUESTION
SBR TRIVIA WINNER 05/06/2013
There's no closed-form solution.
Analytically, what you're looking for is the Lambert W-function.
SBR Founder Join Date: 8/28/2005
So if you do the algebra:
P*k! = λk * e-λ
(P*k!)1/k = λ*e-λ/k
-(1/k)*(P*k!)1/k = -λ/k*e-λ/k
Which would then allow you to use the Lambert W function directly. Specifically:
λ = -k * W[-(1/k)*(P*k!)1/k]
So provided you can find a suitable implementation of the Lambert W, then given any pair of input parameters, k and P,you can come up with arbitrarily precise solutions for λ.
SBR Founder Join Date: 8/28/2005
Thanks guys, I'll take a look at this!
3-QUESTION
SBR TRIVIA WINNER 05/13/2013
3-QUESTION
SBR TRIVIA WINNER 05/16/2013
3-QUESTION
SBR TRIVIA WINNER 05/06/2013