Just listened to this the other day. I really wanted to give some input here – great podcast guys as always, but I had some problems with Marc Alioto's analysis of the hand where he and Jeff were debating whether or not to cbet the AK on J64ss. 

Marc was suggesting – rather strongly – that villain's range for flatting the button was really wide because he was a 23/20 over 69 hands. I would like to disagree with this equally strongly, as I think this is a huge misinterpretation of the villain's HUD stats for a few reasons.

1. If villain is a 23/20 that means he's entering pots as the aggressor almost every time he plays them. The 20% of hands he raises are included in the 23% of hands he plays, which would suggest in simple terms that he only really calls preflop 3% of the time, not 23%. This 3% includes hands where he calls a preflop all-in, limps in, or puts money in the pot after checking a limped big blind. It would be far more instructive in this instance to look at the villain's 'cold call preflop' stat, as that's the one that's really relevant, and his VPIP tells us nothing more that he's been pretty active over the sample of hands we have on him.

2. If I remember correctly, stack sizes in the pot were around 27bb effective to start the hand. Traditional logic would dictate that the fairly short effective stack sizes would narrow the villain's calling range here, and it would be much different to what it would be if it were the first level of the tournament. Additionally, hero was opening from early position, which would narrow the range even further.

3. We only have a sample size of 69 hands on the villain. This is virtually nothing. This is seven and a half orbits of the table. Villain's 23% VPIP indicates that we've seen him play roughly 16 hands at the table. How many of those 16 hands were 27bb effective? How many of those 16 hands did he cold call preflop? How many of those 16 hands was he reacting to a raise from early position? How many of those 16 hands were played against the hero? We know none of this information, and even if we did, we have nowhere near enough of a sample size to infer anything from it.

The gist of what I'm saying is that I think Marc's analysis of the hand was based on some really huge assumptions, and regardless of whether or not cbetting that flop might have been the 'correct' decision, I was quite surprised that none of the TPE pros on the podcast pointed out that we really don't have enough information to use villain's HUD stats to determine his range on this flop. What we do have is the information that he's probably quite active, probably competent given that he's raising more than he's calling preflop, and other reads like nationality, nature of the tournament etc. I would have preferred to see the TPE pros focus on information like this, as I think it's far more relevant than HUD stats in this instance.

This is just my two cents and I'm sure Marc is a far better player than I am, but I spent half of that podcast wishing it was a conversation into which I could insert myself to make that point! 

RonFezBuddy said:

Excellent points.  I think points 1 + 2 are really interesting and I agree with them. Thank you for pointing them out.

 

I disagree midly with point 3.  I understand everything you say about a sample of 69 hands spreading across many situation but I don't think we are implying you take the 16 hands he was involved in and make specific assumptions on specific actions (cold calling preflop for example).   If we have 3000 hands on someone we can probably start making more specifical statistical reads. 10 hands – sure throw it away, but I think in this sample size, you get a general profile for a player over that time that can help your decision making but it should not be a purely quantitative decision about one exact type of action.  

That said, we should not get carried away with data at that sample size.  Use it as another factor in making a decision.  I think your points 1 + 2 are strong enough that we should probably not even be thinking too much about his stats at this point.

I wrote that the other night at like 5am towards the tail end of my session, I think point 3 was just me being a bit of a **** and asking a load of rhetorical questions  Glad to hear my points were valid and ty for the reply RFB.

Well yeah that's obviously true, but you still can't rule anything out, and it's only with extremes like this that you can really make logical inferences. If someone's 29/17 over 100 hands, 14/11 over 10k hands, or indeed 23/20 over 69 hands, there's still room for them to be really good or really bad, and that's going to affect their ranges way more than anything else.