|
|
 Originally Posted by matiusaa
I actually thoughtthat if he had a set he would have bet stronger on the flop, because he wouldn't like to see a club coming.
Please include all of this kind of info in your OP.
This kind of tidbit can change the advice you get on the hand.
Originally Posted by matiusaa
Villian had 89dd, which surprised me a lot.
I am always expecting to see some surprises. 
My statements were cleverly worded as to not overstate my confidence.
This statement holds for T9s as much as for J9s:
"If I see J9 at the end, then I write it off to him running card dead for the past 48 hands. There's no way a player with VPIP of 7 is playing J9 from MP unless it's a misclick or he's just bored."
When he shows up with T9s, I expect the VPIP of 7 to indicate that he's been card dead for a few orbits. It doesn't mean that I treat him like he's playing at 14%, but I will definitely note that he will bet OTF and OTT with a low pair and a draw, and he does big-bet bluffs OTR when he misses his draws.
Originally Posted by matiusaa
Are you sure I can rely on thos stats with confidence?
I used two numbers (VPIP and number of hands) and I extrapolated into a range of ranges.
Somewhere in the 5 - 9% is likely. With all stats, the most likely value is the current value, but the confidence on any single value is 0.
I have no confidence that any stat is exactly what it says it is. I only have confidence when I acknowledge the error of my guess.
The stat is 7
not convincing
The stat is between 6.5 and 7.5
somewhat convincing
The stat is between 6.5 and 7.5 80% of the time, but 10% of the time it's lower, and 10% of the time it's higher
convincing if you can show your work
***
It's about to get a bit nerdy up in here.
If you want the specific confidence intervals, then you asked the right guy.
Assuming Villain's VPIP is 6.3% after 48 hands:
We can say that the "true" value of Villain's VPIP will eventually converge on a value of:
1.6% to 21.5% (99% of the time)
2.1% to 16.8% (95% of the time)
3.3% to 11.6% (75% of the time)
All of these are true at the same time, according to statistical analysis and applying the Wilson Score for the confidence intervals.
Note that the smaller the range we assign to the value, the lower our confidence is in that guess. This is always the case.
We can assume that Villain's VPIP is between 3.3% and 11.6%. We know we will be wrong 25% of the time. 12.5% of the time, Villain's VPIP will eventually converge on a value smaller than 3.3%. Also, 12.5% of the time it will be higher than 11.6%.
We have literally solved mathematically how often we allow ourselves to be wrong. It is our choice which confidence we use. If we pin ourselves down to demanding a very tight range, then we choose to act with low confidence.
However, what if a Villain was showing a VPIP of 6.3 after 240 hands? Now we have a lot more data backing up the value of 6.3. What impact does that have on our confidence?
We can say that the "true" value of Villain's VPIP will eventually converge on a value of:
3.3% to 11.6% (99% of the time)
3.8% to 10.1% (95% of the time)
4.7% to 8.3% (75% of the time)
This time we can assume that 0.5% of the time, Villain's VPIP will eventually converge on a value smaller than 3.3%.
Also, 0.5% of the time it will be higher than 11.6%.
So the fact that we have collected more data gives us higher confidence with the same error bars. OR it can give us thinner error bars at the same confidence.
|
Reply With Quote