|
|
#1
07-07-2005, 03:24 PM
|
|||
|
|||
Statisticaly Significant Results
A Question for those of you with a more analytically bent mind. If I bet a consistent unit over how many wagers would I have to show a total profit and how much of one to tell if I am good at this? Let's say my unit is $100. if after 300 wagers I am up $2,000 or 20 units, is that statistically significant? If so, or if not by what process or formula did you reach that conclusion?
Thank you in advance, Luke 
|
|
#2
07-07-2005, 03:29 PM
|
|||
|
|||
|
Re: Statisticaly Significant Results
This is just a guess, but the standard deviation of a binomial distribution (which I think this is, sort of), is SQRT(np(1-p)) where n = sample size and p is the probability of winning, which I guess we can assume is 50%, since we're working in units. For n = 300, p = 0.5, you get a STDEV of 8.66 units. So if you're up 20 units after 300 wagers, you can be about 99% sure you're a winning bettor.
|
|
#3
07-07-2005, 04:37 PM
|
|||
|
|||
|
Re: Statisticaly Significant Results
That makes sense, my college stats classes are coming back to me...So find the STDEV and compare it to a bell curve, if I'm 3+ STDEV's than it is 99% likely I'm a winner in the long run. Between 2-3 STDEV's would be 95-99% and 1-2 STDEV's would be 67-95% and between 0-1 STDEV's would be 50-67% if I remember correctly. Thanks ctv!
Any other thoughts on this?
|
 |
|
|
Linear Mode
