Unfortunately this is flawed because of the issue of self-selection.

The players who win are more likely to stay at a table and at a site for longer than those who are losing. And if you aren't capturing the majority of hands played on a site then you don't know what happens next.

There is another issue. Imagine 2 sorts of player - one rebuys at any table where he gets stacked.The other decides from experience that with his image shot and possibly being outplayed he would prefer to open a fresh table (usually one you aren't logging). The second player will often look like he is one of the <100 hand donks to you and the first won't - but they may well each be equal winners and losers

See the attempt to work around this problem in the PT autorate article as I compare the 'persistence' of player types (ie their differing drop out rates) over different nos of hands logged.

http://www.andymcnish.btinternet.co.uk/newauto.htm

I supect you are quite right in your theory btw, I just dont think your numbers 'prove' it as well as you think.

Hey Ex.

Read your work on drop out rates while sitting in an Airport one day way back when you wrote it.

Not being a statistician, I don't know what degree of certainty my sample provides (or whatever those guys call it), but I would expect other similar samples to show similar results.

Not all that worried about missing that guy that changes tables though. Frequently, I'm playing every 100NL game they have running while I'm there.

Self selection is kinda the point though. They come, they lose, they weed themselves out.

The nastiest example I saw was someone who dropped 3 1/2 buy-ins in 4 rounds ... ne'ery to be seen again. Many folks I've seen show up, play 10, 5, even 1 hand, lose all their chips ... and that's the last I ever see of them.

Another interesting note was how the winner to loser ratio really didn't shift all that much. 1 to 1 as you get longer term (probably some kind of zero sum math going on there), but only 1 to 2 for even the shortest of players. I would have expected a larger shift.

Anyway, C ya round Ex.

Hell no...switched to 6max a couple months ago and that's simply not profittable there.