Quote:
Originally Posted by wayneking7
Yeah that was me unfortunately
Can you run a sim on my results? But that about half 6m hypers and half F50, so std dev obv diff for the 2 but im sure we can get an approx. figure
Sorry, but it only works for an actual run of hands and I don't have a parser written for anything else. If anybody is interested then here's what I did:
The algorithm to calculate EV_diff (ie: difference/gap between the red and green line) is this:
Code:
EV_diff = 0
For each all-in:
{
Calculate the weighted average $equity of the possible all-in outcomes, W
Calculate the actual $equity of the outcome that occurred for you, A
EV_diff = EV_diff + (A-W)
}
You then end up with a number which is the amount of $equity you ran above EV (positive) or below EV (negative). This is the amount that is "added on" to the green line-in HEM to get the red-line (if you divide this by 0.1 for 10-man or 0.1667 for 6-man, etc you get the number of buy-ins you are running above/below if needed...).
A monte-carlo algorithm to work out how bad you are running is very similar:
Code:
Num_worse = 0
For N simulations:
{
EV_diff_monte_carlo=0
For each all-in:
{
Calculate the weighted average $equity of the possible all-in outcomes, W
Choose one of the possible outcomes for the all-in randomly (ie: Sampling from a discrete distribution / biased roulette wheel sampling, etc), R
Calculate the $equity of the outcome R chosen above, A
EV_diff_monte_carlo = EV_diff_monte_carlo + (A-W)
}
if EV_diff_monte_carlo<EV_diff
Num_worse = Num_worse + 1
}
Percentile = 100*(Num_worse/N)
So you now have a number that tells you the % of times you'd expect to run worse for that particular section of games/all-ins (eg: Percentile = 1% for a months worth of data then this [or worse] is only expected about 1 in every 8.3 years, etc).
You can also calculate the SD of the EV_diff_monte_carlo bit and use the area under a Gaussian, but I've no idea how you can get/estimate this from the payout, game-type, etc (as mentioned above it could even be player/strategy dependant as to the types of outcome distribution you'll see...).
Maybe somebody with some time on their hands could write this to work with the stars hypers as it is very helpful (for you sanity!) to be able to gauge just how extreme a bad run is in terms of rarity (this is what I suggested to bumpking a few years back when he had a really bad run and I think he can vouch for it helping him).
Juk
Last edited by jukofyork; 03-15-2013 at 02:19 PM.