[jukofyork]

[ QUOTE ]
[ QUOTE ]

If you want to try to qualify the results in %s, then one thing you can try is to first convert each of the K/S rankings into a percentage. Then every-time you see a player call your push in a certain situation that you are examining, write down the %-rank. Finally after getting lots of these find the mean and double it. EG:

Average donk / called SB push with effective stack of 8BB / 1st SB->BB push : 10%, 8%, 9%, 20%, 13%, 5%, 2%, 8%, ...

(10+8+9+20+13+5+2+8)/8=9.375%, 9.375*2 = 18.75%

So you can now assume that if he is using the K/S ranking system and only calling the top N% vs you, the cut-off he is using is about top 19% of hands (this should also give you a rough and ready idea about your fold equity too). You need quite alot of data to do this though, but you should get some idea (perhaps +/- 5%) of calling range %'s in different situations by trying it out.

Juk [img]/images/graemlins/smile.gif[/img]

[/ QUOTE ]


Why double it?

[/ QUOTE ]
Because the average hand you observe them call with is in the very center of the distribution of hands they call. So if you see them calling with a top 5% hand on average, then their cut-off is top 10%, and if you see them call with a top 50% hand on average, their cut-off is top 100%, etc.

This assumes they use a simple binary threshold (call/don't call), but in practice (and especially if you are grouping player types together) the histogram would have an a S-shaped curve centered on the mean which signifies that sometimes they would call and sometimes they would not (no SNGPT like tool can handle these frequency distributions yet anyway...).

Juk [img]/images/graemlins/smile.gif[/img]