Quote:
Originally Posted by leolauzon
But I guess your stance is that we can't discuss live tournament players because none of them has a big enough sample size?
You need much less sample size in live than online, especially when it comes to players (like Hellmuth) who are superior at reading tells and also superior at influencing people's behaviors and emotions with words, demeanor and motions.
Tells and the ability to influence people's emotions and therefore behaviors to one's liking can lower variance considerably. For instance, let's take the extreme example of me noticing that a particular player picks his nose 100% of the time he over bet bluffs all in when the stacks are super deep. Then there would be zero variance in this particular maximum profit situation event for me. This is an extreme example. But Hellmuth probably sees tons of lesser tells that if accumulated can lower variance over time.
Online, the only info Hellmuth would have would be frequencies and ranges in terms of hands voluntarily played, c-betting, double or tripling barreling, etc. But live, a highly perceptive tell reader like him could just find a one tell tat could override all historical frequencies.
Now, you accumulate all of these tells and many instances in which Hellmuth would be "setting up" many opponents with conditioning plays and language, and you have cut down variance considerably. Hence, cutting down sample size to perhaps as little as a tenth compared to if he just depended on frequencies and ranges that show up in HUD.
Remember, poker is a game of imperfect information. A single instance of a person picking his nose 100% of the time that he pure bluff bets $1,000 into a $100 pot on the river turns it into chess. If I played chess against Kasparov, you wouldn't even need an entire game as the sample size to conclude that I stand no chance in the long run. In fact, just the first ten moves would be enough of a sample size to have 100% certainty that Kasparov has infinite EV over me.
In short, the people that say Hellmuth does not have a big enough sample size are idiot nerds who think that ranges and frequencies that are recorded in HUD are the entire universe. They have been infected by their own cognitive biases and so have their opinions.