Statistical Analysis

[Crazy-Ivan1978]

Just Curious purely for my own sake. Anyone able to give me some real meaning to these statistics for a newbie?


Statistics for 213 Hands
Street Saw Saw/Total
Flop 122 57%
Turn 89 42%
River 80 38%
Showdown 68 32%
Street Won Won/Saw Won/Total
Pre-flop 7 3% 3%
Flop 14 11% 7%
Turn 4 4% 2%
River 7 9% 3%
Showdown 34 50% 16%

[Crazy-Ivan1978]

Statistics for 248 Hands
Street Saw Saw/Total
Flop 147 59%
Turn 107 43%
River 97 39%
Showdown 83 33%
Street Won Won/Saw Won/Total
Pre-flop 7 3% 3%
Flop 18 12% 7%
Turn 4 4% 2%
River 8 8% 3%
Showdown 43 52% 17%


This is how i ended for the night. Started with $1000, left with 49,000 in play money.

[Muck]

Please read the following because it's important:

1) Analysis needs to be done over 10’s or even 100’s off THOUSHANDS of hands.

2) Play money is so different from real money that the analysis is meaningless.

[suitedaces84]

1) Analysis needs to be done over 10’s or even 100’s off THOUSHANDS of hands.

Not really. Most statistics converge relatively quickly (VPIP, PFR, att. to steal, AFq, etc). It's just win rate that you need to wait a long time for.

[Muck]

Really? But surely such a small set is impacted by the type of players at that session. E.g. you may see a lot to rivers but that could be down to a passive table that only likes to bet early.

Do you mean that two hundred hands is enough, or that thousands are just unnecessary? Or both?

[Jauron]

That preflop % is outrageous. It looks like OP calls too much, and is not very aggressive. It's hard to win if you are constantly showing down hands, especially when you play as many as you do.

[Muck]

I think with play money a high flop % is okay, if your strategy is to play cheap, hit strong then overbet to make it worth it.

[Crazy-Ivan1978]

It's weird because in play money games i expect people to call me more with bullshit but i still end up winning my hands.

In live games its quite reverse. I am pretty stone cold when it comes to playing poker and when i play with real money im friends with no one. Just trying to beat out everyone else honestly except in live games i just end up taking a beating.

I know the basics and i never go into a hand without something real in my hole cards so yeah i am just trying to get a feel whether its me or its just bad luck.

As a beginner i don't know all the statistical odds yet (I've only been playing 4 months) however i do feel i have a pretty good natural instinct about the strength of my hands.

Just trying to see if i should focus more online and build my strength or keep entering live tournaments in an attempt to build my knowledge of players.

Thanks guys in advance for all the advice.

[suitedaces84]

Really? But surely such a small set is impacted by the type of players at that session. E.g. you may see a lot to rivers but that could be down to a passive table that only likes to bet early.

Do you mean that two hundred hands is enough, or that thousands are just unnecessary? Or both?

Most of the stats used by the OP are binomial. The standard deviation is sqrt(p*(1-p)/n).

For example, the saw flop has a SD of ~sqrt(0.59*0.41/248) = 0.034.

[Muck]

It's weird because in play money games i expect people to call me more with bullshit but i still end up winning my hands.

In live games its quite reverse. I am pretty stone cold when it comes to playing poker and when i play with real money im friends with no one. Just trying to beat out everyone else honestly except in live games i just end up taking a beating.

I know the basics and i never go into a hand without something real in my hole cards so yeah i am just trying to get a feel whether its me or its just bad luck.

As a beginner i don't know all the statistical odds yet (I've only been playing 4 months) however i do feel i have a pretty good natural instinct about the strength of my hands.

Just trying to see if i should focus more online and build my strength or keep entering live tournaments in an attempt to build my knowledge of players.

Thanks guys in advance for all the advice.

Post example hands. I find it’s an easier focal point for discussion. Do a good mix of ones you won, ones you lost, ones you folded. If they’re from a live game try and add as much detail as possible.

Given the choice I think you’re better off learning on-line. The games are cheaper so you get more experience-hours for your money.

Really? But surely such a small set is impacted by the type of players at that session. E.g. you may see a lot to rivers but that could be down to a passive table that only likes to bet early.

Do you mean that two hundred hands is enough, or that thousands are just unnecessary? Or both?

Most of the stats used by the OP are binomial. The standard deviation is sqrt(p*(1-p)/n).

For example, the saw flop has a SD of ~sqrt(0.59*0.41/248) = 0.034.

I took mechanics rather than statistics at school so I’m not familiar with this, however I’ve been Wiki’ing though your post and now have a better idea of what you’re explaining…however I’m still finding it a bit of a head f**k.

[suitedaces84]

That's the awesome thing about wikipedia. It's jibberish unless you already know it, in which case it's useless.

Anyway, here's my explanation:

A risidual is the difference between the mean and the result. Variance is the expected value of the risidual squared. Standard deviation is the square root of variance.

For example, if p = 0.3 then 0.3 is the mean.

70% of the time the risidual will be -0.3; 30% of the time the risidual will be 0.7. Note that the average risidual is 0.

70% of the time the risidual squared will be 0.09; 30% of the time the risidual squared will be 0.49.

var = 0.7*0.09+0.3*0.49 = 0.21 = 0.3*0.7

More generally:

The risidual will be -p with probability 1-p; the risidual will be 1-p with probability p.

The risidual squared will be p^2 with probability 1-p; the risidual squared will be (1-p)^2 with probability p.

var = p^2*(1-p) + (1-p)^2*p

= [p^2*(1-p)] + [1-2p+p^2]*p

= [p^2-p^3] + [p-2p^2+p^3]

= p - p^2

= p*(1-p)

This is for a single trial (n=1).

Both variance and expecation add--standard deviation does not. It's not tough to see that variance and standard deviation cannot both add--think Pythagorean theorem. If a^2 + b^2 = c^2, then a + b != c (unless a = 0 or b = 0).

For variance to add the events need to be independent--if they're not independent covariance term(s) need to be added. The covariance of independent events is 0.

In hold'em (unlike 7 card stud) the events are not quite independent. You're more likely to play a hand in LP than in EP. The covariance term will actually be negative. There's slightly less variance in flops seen % than there my first calculation implied.

[marios_521]

I think with play money a high flop % is okay, if your strategy is to play cheap, hit strong then overbet to make it worth it.

+1

[marios_521]

Really? But surely such a small set is impacted by the type of players at that session. E.g. you may see a lot to rivers but that could be down to a passive table that only likes to bet early.

Do you mean that two hundred hands is enough, or that thousands are just unnecessary? Or both?

Most of the stats used by the OP are binomial. The standard deviation is sqrt(p*(1-p)/n).

For example, the saw flop has a SD of ~sqrt(0.59*0.41/248) = 0.034.

This is for the Normal Distribution, but anyway can you explain your thinking? (what is OP?)

EDIT: WE POSTED AT THE SAME TIME... I 'LL READ YOUR POST/

[Muck]

(what is OP?)

Original Post, i.e. the first post in the thread.

...I think anyway.