Thanks to all for the answers. My apologies for not framing my original post correctly, it's not about my play or my sample, I am referring to general play.

Let me see if I can put the point over in a completely different way.

When playing poker, we face choices:

Freflop: Fold, Call, Small Raise, Big Raise.

If not folded then proceed to after flop action.

Flop: Fold, Check/Call, Small Raise. Big Raise.

If not folded, then proceed to after turn action:

Turn: Fold, Check/Call, Small Raise. Big Raise.

If not folded, then proceed to after river action:

River: Fold, Check/Call, Small Raise. Big Raise.

Either win or lose or tie.

So, the decision tree for playing a hand will always be the same. Now, here is where I want to examine the playing theory.

__Higher level games: (ie higher level of understanding)__
The above decision tree also has some rules applied.

Pot odds and value calculations are used, such that if it's not profitable to make a call, then the hand is folded. Therefore, outcomes are predictable and it's easy to see how some hands are lost through variance, but as long as the game is played according to the rules, then over time you should expect plays to be profitable. There is a mathmatical model that shows to be true in the long run.

If we were to plot the decision tree for each possible starting hand type, we would be able to statistically show that the conventionally accepted stronger hands are more profitable. We don't need to look at anyone's hand history, the theory holds mathmatically valid. If we analysed millions of hands at this level, we would expect the analysis to show that in reality, the long term value of hands matched the theoretical model. We all accept this as general poker theory.

OK, now the big question:

__Lower level games (ie micro stakes and lower skill level/level of understanding)__
Now we have the same decision tree as above, however we remove the rule that states your plays are based on odds, expectations and values. Put simply, people call when they shouldn't and so you run the risk of getting suckered more.

Now, we see many more hands that reach the end of the decision tree, because now they are not being folded as they are not subject to the rules stating that hands that are not mathmatically profitable to call should be folded.

We don't have a mathmatical model now. We have a statistical one that can only be demonstrated by data mining a hand history. The more hands analysed, the more accurate the results.

By data mining a large sample set (and I mean LARGE), you would come up with some interesting statistics. From that you would be abkle to derive what hands play best and in what way - or, which opening hands, and which decision tree, was the most profitable.

For example, this method may show that, over a billion hands, J10s was the most valuable starting hand with the highest frequency win rate, and Q10 was the hand with the biggest pot wins. (These are examples only).

Based on these results, your strategy would be to play the decision trees of the hands that do best at this level.

We can't determine that mathematically because the hands aren't played according to mathmatical rules.

In my experiences, there are players that we all label as lucky fish, maniacs, river rats, and so on, because we lose good hands to what seems to be bad play. But...

What if these guys are actually just naturally playing the most profitable decision trees ?

Until some maths whiz does a study, then we can't really know.

(For those that doubt the theory, look into data mining used by financial institutions. Here they specialise in running queries against huge data sets to find trends that conventional maths won't spot. It's a sound concept).

Thanks for listening....