chi-test

[rumply]

Hi all, 1st post here

have been reading with interest recent posts about the chi-test & would like someone to confirm my calcs, as they seem a bit too good to be true -

btw i bet place only & these are figures for the last 5 months-

average dividend $2.33
highest div $7.8
lowest div $1.1
pot 16.9%
runners 1111
places 557
loosers 554
% avg winners 50.1%
expected no of winners 414
expected no of loosers 697
chitest score - 0.00000000000000078

if anyone would care to offer their thoughts on the above results it would most appreciated

& also regards the average dividend - i have used unitab divs to calculate the figure above, but i bet solely with Betfair & on average achieve a better divy, esp with my longer priced selections
<table x:str="" style="border-collapse: collapse; width: 159pt;" border="0" cellpadding="0" cellspacing="0" width="212"><tbody><tr><td class="xl25" style="height: 12.75pt; width: 159pt;" x:num="7.7963700773228105E-19" height="17" width="212">
</td></tr></tbody><tbody><tr style="height: 12.75pt;" height="17"></tr></tbody></table>

[AngryPixie]

... although it's potentially a bit complicated due to the differences in market% due to field size when betting for the place. Ideally you'd be recording the market% of each race and normalise each selection individually, then aggregate the totals instead of using the average SP.

What the result is saying is that to this point it's very unlikely that what you've observed is due to luck. It's a tool, not a guarantee that things will continue along the same path, and assumes that the SP is an accurate indication of a selections chance. ;)

[jfc]

I suspect very few people here could confidently perform a CHITEST.

For this Binomial Confidence Limits are much easier.

http://statpages.org/confint.html

557/1111 produces

47.15% as a lower bound.

For an average dividend of $2.33

Your expected POT is 9.9%

Most people would not sneeze at a POT that high.

[Chrome Prince]

rumply,

According to my benchmarks, that system is phenomenal, especially for 5 months results containing that many bets.

You are relying on 24 winners for the profit - a massive plus.

Get another 5 months under your belt, and I think you could successfully conclude it's a good'un.
The prognosis is for a long healthy life after the pathology results came back ;)

[Mark]

Would anything survive the Hay Chee test??? Let's win win win.

[Chrome Prince]

Maybe the Tai Chi test?

;)

[Chuck]

Would anything survive the Hay Chee test??? Let's win win win.

Just got to visualise it mate

[rumply]

rumply,

According to my benchmarks, that system is phenomenal, especially for 5 months results containing that many bets.

You are relying on 24 winners for the profit - a massive plus.

Get another 5 months under your belt, and I think you could successfully conclude it's a good'un.
The prognosis is for a long healthy life after the pathology results came back ;)hit my funny bone there Crome Prince, as 4 weeks ago i underwent my 1st Colonoscopy & recently went back to the surgeon for the results - a couple of polyps but thankfuly pathology all clear:) (too all you 45&> yolds out there - if you havent had one yet, you dont know what your missing!, ahh the joys of life & i get to do it all again in a few years...)

sorry bit too much personal info for a 2nd post i know, back to the punt

my selections are based on self generated ratings & then filtered for price, late market movements & a couple others that are no more than common sense, & as i hit the 1000+ bet mark thought it was time for some outside opinion.

thanks for all the replies