Best way to play statistically?
 

This may be purely theoretical, but I'm wondering what statistics translate into the best BB/100 over time. Obviously successful players can have wildly different approaches, and therefore wildly different stats, but I'm betting that over time we could get some pretty concrete numbers. Take for example, the following 5 stats:

1 VPIP
2 PFR
3 Aggression Factor
4 Attempt to Steal Blinds
5 Went to Showdown

I just suggested these, because they are what I look at most of the time. What are the optimal numbers for these?

Also, it would be cool if there was some way for people to upload their PT stats to a big anonymous database, so that we could have all of this stuff in black and white.

Lastly, feel free to add stats that you may think are more important.

Re: Best way to play statistically?
 

I have had great success with

1. Blah
2. Meow
3. Chow
4. Lots
5. Lots

Re: Best way to play statistically?
 

Check out the bots thread.

Whatever stats the bots use must be approx. the best stats right? But for human players the stats are not quite as meaningful, thus I don't expect you to get many responses.

The key is when to check - raise, be highly aggressive, passive, for the most money. Generally, yes the bots are correct. But for each unique hand, you must assume they are not choosing the best longterm strategy every time.

In my opinion, poker at party poker had a thin winning margin (some say 95% of players are long time losers). If you assume that you can win 5BB/100 (this is fairly high) at $3/$6...then I think that means $30 but it might be $15. That means the you only need to make 5 mistakes (missed bets or bad bluffs) to earn $0/100. Do bots make 5 mistakes per 100 hands? Well, from what we know, they have a win rate of around 1.5BB/100, so they are making about 3.5 mistakes per 100.

Regards.

Re: Best way to play statistically?
 

poker is not that easy. Optimal stats for a tight table and a loose/passive table and loose/aggressive table are going to be significantly different. unless your database was for a specific limit at a specific site, you probably won't get anything useful.

Re: Best way to play statistically?
 

Now these stats will work for any poker game, at any limit/structure, with any number of players in the game HU-20 person fr, & with any player type (what is there like 4 types now?).

1 30
2 10
3 1.50
4 .75
5 52.5

OBV! Code it up & make a fortune, we've solved the great poker math problem.

Re: Best way to play statistically?
 

You probably could get a quite narrow "ideal" range for limit, but you would have to make certain assumptions about the relative looseness and aggression of the opposition.

Not that this helps much, because even being in this ideal range (even assuming your base assumptions are correct) does not ensure success. Yes you need to raise, and be aggressive etc. but in the right place at the right time. That's where the skill/experience comes in.

Re: Best way to play statistically?
 

Statistically speaking, scientists now believe, playing well is better than playing badly.

Hope that helps.

*note: the above does not apply to Ultimate Bet, which is rigged*

Re: Best way to play statistically?
 

It's like kung fu. Here's my shot.

LAG style
1 VPIP - 60%
2 PFR - 25%
3 Aggression Factor - 5
4 Attempt to Steal Blinds - 70%
5 Went to Showdown - 30%
6 Win at Showdown - 100%

TAG style
1 VPIP - 10%
2 PFR - 5%
3 Aggression Factor - 7
4 Attempt to Steal Blinds - 20%
5 Went to Showdown - 10%
6 Win at Showdown - 100%

FISH Style
1 VPIP - 80%
2 PFR - 40%
3 Aggression Factor - 2
4 Attempt to Steal Blinds - 50%
5 Went to Showdown - 100%
6 Win at Showdown - 20%

Drunk Friday night style
1 VPIP - 100%
2 PFR - 100%
3 Aggression Factor - oo
4 Attempt to Steal Blinds - 100%
5 Went to Showdown - 100%
6 Win at Showdown - 10% (16% if 6-ring)

Re: Best way to play statistically?
 

You might find this thread useful.

Re: Best way to play statistically?
 

That's exactly what I was looking for. Thanks Sniper.