How big is your edge?

[Big_Jim]

[ QUOTE ]
[ QUOTE ]
To elaborate, some very high EV situations can have a ton of variance, and result in very long losing streaks.

Easy/obvious example:
You are getting paid a billion to one on a million to one shot.

Your EV is huge, but you can have very very long streaks where you are losing.

While obviously extreme, most poker situations concern probabilities where you are getting large payoffs for situations that don't happen frequently, as opposed to the coinflip.

[/ QUOTE ]

Not necessarily true. I'm flipping coins all day.

For example, I raise and button three bets. I push all-in for 100bb. He calls with his range -- JJ+,AK.

My pushing range is QQ+, AK. My expected equity for this flip is 52.92%

[/ QUOTE ]
And how many times a day does this happen? How many hands do you play?

[_dave_]

[ QUOTE ]

Any chance you could tell us how to get the total amt bet number from our PT database without knowing anything about coding?


[/ QUOTE ]

Sure [img]/images/graemlins/smile.gif[/img]

do "Start -> Programs - > Postgresql 8.x -> pgAdmin III"

double-click your database on the left of pgAdmin to connect to it. The red X should disappear.

In pgAdmin, do "Tools -> Query Tool"

paste the SQL code in the top part of the window that pops up, edit screenname as desired, do "Query -> Execute".

You should get hands, total_bet, total_won, "edge" and profit returned for whatever screenname you edit in to the SQL code you copy/pasted [img]/images/graemlins/smile.gif[/img]

[Lyric]

[ QUOTE ]
And how many times a day does this happen? How many hands do you play?

[/ QUOTE ]

Yep, just an example. As I said before, it's pretty complicated, myriad conditions apply, and we'd really need to see a new stat from PT (or use code to pull it out) to get a good handle on it.

[soah]

If you're playing in games where you're frequently playing huge pots where you aren't sure whether you're ahead or when you're drawing, then your variance will be huge. This is nothing new.

On the other hand, if you're playing against weak players who don't disguise their hands and play terrible in general, you're generally going to have a huge edge when you get much money in the pot. Variance in these games for a decent hand reader is quite small.

Even though your percent edge is much bigger when you get your money in the pot in the second game, the winrates can still end up similar because against the weaker/passive opponents the pot sizes will be much smaller on average.

[Big_Jim]

The EV of your 50.5% edge player is:
1*0.505 - 1*0.495 = 0.01

If I run 10,000 simulations, with 32000 trials each:

9619/10000 winners. 0.9619%

For a reference point, I ran 10,000 simulations of 32000 trials on the following scenario:
9*0.101 - 1*0.899 = 0.01

7254/10000 winners. 0.7254%

I'll post the sims, upon request.

[Big_Jim]

[ QUOTE ]
[ QUOTE ]
And how many times a day does this happen? How many hands do you play?

[/ QUOTE ]

Yep, just an example. As I said before, it's pretty complicated, myriad conditions apply, and we'd really need to see a new stat from PT (or use code to pull it out) to get a good handle on it.

[/ QUOTE ]
Perhaps I should have read the thread? ah well

[KRANTZ]

[ QUOTE ]
mines so [censored] big

[/ QUOTE ]

[Lyric]

[ QUOTE ]
The EV of your 50.5% edge player is:
1*0.505 - 1*0.495 = 0.01

If I run 10,000 simulations, with 32000 trials each:

9619/10000 winners. 0.9619%

For a reference point, I ran 10,000 simulations of 32000 trials on the following scenario:
9*0.101 - 1*0.899 = 0.01

7254/10000 winners. 0.7254%

I'll post the sims, upon request.

[/ QUOTE ]

I'm not quite sure what this means. I think it means that our 50.5% coin flipper, if flipping for $100, would be expected to win (0.505-0.495)*$100 = $1 per flip. Yes?

If someone is "flipping" AA vs KK all-in pre-flop for $100 he is expected to win (.80-.20)*$100 = $60, yes?

So the "edge" for the AA vs KK player is 60%. The edge for our coin flipper is 1%. Yes?

[Lyric]

Okay, so I got the SQL request code to work for my database, which has 295K 5/10 NL 6max hands in it. My edge for these hands is 1.78%, after rake. This means for every 1K I put in the pot, I expect to win $17.80. Awsome. (Running at 2.2ptbb/100 for these hands)

Here is the expected variance for a 1.78% edge over 32K hands. None of these outcomes would be anything but completely ordinary for my edge/skill level. This is pretty sick considering 32K hands is about 10 months of playing live poker full time:


Note: The edge percentage is calculated using the total amount "bet," and but not necessarily put at risk. So it's not accurate in cases where you push all-in for $1,000 and your opponent only has $63 in front of him. In that case you'd still be counted as "risking" $1,000 when in fact you only risked $63. Not perfect, but a pretty fancy guess.

[soah]

[ QUOTE ]
This is pretty sick considering 32K hands is about 10 months of playing live poker full time

[/ QUOTE ]

but your edge in live games is much larger, and the variance is much lower