[brettd]

Hi guys, just wondering what common analytics you guys use to examine your results? I.e; medians, st dev's, significnce tests, etc....

[Wrecktangle]

The metrics I use depend upon the technique, but overall I use the standards: win% & units won per season, mean squared error.

[brettd]

Can you some provide some more info please wreck? Right now, i've got the obvious stuff for each of my betting areas. Amounts outlayed, ROI%, number of betsm and strike rate %.

I want to essentially determine how statistically signifcant my ROI% results are, and with what % certainty I can expect these results going forward.

[u21c3f6]

Hi guys, just wondering what common analytics you guys use to examine your results? I.e; medians, st dev's, significnce tests, etc....

Here’s how I look at this.

When I want to know if my results are “valid”, I check my results against what could happen by random chance. First, I always group my results by the type of wager. For example, I don’t mix ML and side wagers or in other words, I don’t want to mix lines of 100 and 200 etc together.

I will assume I am testing a sample of 400 side wagers with an average no-vig line of 100. First I calculate the std dev (Sq Rt (400*.5*.5)) = 10. By chance, on average, one should have 200 winners. One s.d. would be 190 to 210 winners, 2 s.d. 180 to 220 winners etc. I like my results to be greater than +2 s.d. and ideally greater than +2.5 s.d.

Typically, the larger the sample, the greater the accuracy.

Joe.

[brettd]

Ok so how would I apply it to this? I have 300 wagers at various prices recorded. My strike rate for those 300 wagers is 17.88%. I know for a fact that long term SR% is 35% for this betting segment (recorded over 10,000 bets).

What's the likelihood of running at 17.88% given a long term strike rate of 35%?