Variance Simulation

I've only recently found out about this site and some of the results that it shows are quite depressing. It basically shows that poker is more of a gamble that it actually is a skill game.

For example, here is a graph of 900 runs, the max/min of a -10bb/100 through 100k hands at HU.



As you can see, someone who is pretty big loser can still show to be a winner over 100k.

What this means in my eyes is that if you are planning to play poker for a living, unless you are putting in MASSIVE volume or playing a lower variance gametype you will never actually know if you're winning player.

seen something similar in a old thread. nothing new, poker is party a game of luck and i doubt it ever balances out.

i m 150+ bis under ev after 2M hands and its getting worse every day

poker is bullshit

/whine

A lot of people don't understand what EV actually mean.. EXPECTED value. These graphs don't show an accurate story. For example whenever you go all in with a set, you generally have the EV of 60%-95% of a buy in, when you go all in preflop with KK you generally have the EV of ~66% of a buy in etc..

A simple example: You get 88 all in on K89hh and run into bigger sets 2x, then you go all in with KK and suckout vs AA two times are you running good or bad? Are you playing / running well or poorly. According to an EV graph it will say you are running above EV despite being even, but when you calculate your expected rate of return with those hands based on the equity on opponents range you have an expectectation of almost +3 buy ins, but according to an EV graph you are even, and have an EV of -2 buy ins

On August 25 2010 13:10 TalentedTom wrote: A lot of people don't understand what EV actually mean.. EXPECTED value. These graphs don't show an accurate story. For example whenever you go all in with a set, you generally have the EV of 60%-95% of a buy in, when you go all in preflop with KK you generally have the EV of ~66% of a buy in etc.. A simple example: You get 88 all in on K89hh and run into bigger sets 2x, then you go all in with KK and suckout vs AA two times are you running good or bad? Are you playing / running well or poorly. According to an EV graph it will say you are running above EV despite being even, but when you calculate your expected rate of return with those hands based on the equity on opponents range you have an expectectation of almost +3 buy ins, but according to an EV graph you are even, and have an EV of -2 buy ins

if you were just responding to dogmeat then nvm but what you're talking about and what the op posted are completely different things.

On August 25 2010 13:10 TalentedTom wrote: A lot of people don't understand what EV actually mean.. EXPECTED value. These graphs don't show an accurate story. For example whenever you go all in with a set, you generally have the EV of 60%-95% of a buy in, when you go all in preflop with KK you generally have the EV of ~66% of a buy in etc.. A simple example: You get 88 all in on K89hh and run into bigger sets 2x, then you go all in with KK and suckout vs AA two times are you running good or bad? Are you playing / running well or poorly. According to an EV graph it will say you are running above EV despite being even, but when you calculate your expected rate of return with those hands based on the equity on opponents range you have an expectectation of almost +3 buy ins, but according to an EV graph you are even, and have an EV of -2 buy ins

I agree completely with your analysis of EV graphs and think they're completely useless, however what i posted is not all-in EV. I was also going to say the same to dogmeat but didn't want to this thread to degrade into a derpfest of whether EV graphs actually mean anything or not.

edit: I really liked your post actually and im going to link to it whenever someone posts and ev graph an starts whining

it's best to not even think about shit like this, keep your mind clean and open for the game
this just makes me depressive

I think understanding variance is crucial, this has helped me a lot dealing with down and upswings and helps put things into perspective. As a poker player you need to be able to accept that variance can affect you in massive ways and this has only helped improve my game imo.

variance is ez

On August 25 2010 10:39 fenner wrote: I've only recently found out about this site and some of the results that it shows are quite depressing. It basically shows that poker is more of a gamble that it actually is a skill game. For example, here is a graph of 900 runs, the max/min of a -10bb/100 through 100k hands at HU. As you can see, someone who is pretty big loser can still show to be a winner over 100k. What this means in my eyes is that if you are planning to play poker for a living, unless you are putting in MASSIVE volume or playing a lower variance gametype you will never actually know if you're winning player.

Although it's true that out of 900 runs the hottest run would be actually profitable over 100k hands when his true winrate is -10bb/100, he had to be running very hot. So maybe his all-in EV is going to be sick above expectation, his sets flopped are high, high pocket pairs much more frequent than expected, or most likely a combination of all these things.
You can figure out over 100k hands whether you're a profitable player or not. The super lucky guy running above everyone else just can stay in denial for longer later on when his luck starts normalizing.
I know it's really only on my downswings that I really take the effort to improve and thoroughly look over my game. If I always ran as expected I'd just be satisfied and probably get lazy.

On August 25 2010 15:21 jchysk wrote: Show nested quote + On August 25 2010 10:39 fenner wrote: I've only recently found out about this site and some of the results that it shows are quite depressing. It basically shows that poker is more of a gamble that it actually is a skill game. For example, here is a graph of 900 runs, the max/min of a -10bb/100 through 100k hands at HU. As you can see, someone who is pretty big loser can still show to be a winner over 100k. What this means in my eyes is that if you are planning to play poker for a living, unless you are putting in MASSIVE volume or playing a lower variance gametype you will never actually know if you're winning player. Although it's true that out of 900 runs the hottest run would be actually profitable over 100k hands when his true winrate is -10bb/100, he had to be running very hot. So maybe his all-in EV is going to be sick above expectation, his sets flopped are high, high pocket pairs much more frequent than expected, or most likely a combination of all these things. You can figure out over 100k hands whether you're a profitable player or not. The super lucky guy running above everyone else just can stay in denial for longer later on when his luck starts normalizing. I know it's really only on my downswings that I really take the effort to improve and thoroughly look over my game. If I always ran as expected I'd just be satisfied and probably get lazy.

LOL

Of course the graph looks like that. You posted max/min for NINE HUNDRED runs. If you do 20,000 trials then your max will be even higher. If you manage to hit that then, yes, you did get very very lucky.

It takes at least 1/2 mil hands to get the idea (assuming 6max and typical SD). But the truth is you can never be sure - risk of ruin can never (theoretically) be zero.

On August 25 2010 13:10 TalentedTom wrote: A lot of people don't understand what EV actually mean.. EXPECTED value. These graphs don't show an accurate story. For example whenever you go all in with a set, you generally have the EV of 60%-95% of a buy in, when you go all in preflop with KK you generally have the EV of ~66% of a buy in etc.. A simple example: You get 88 all in on K89hh and run into bigger sets 2x, then you go all in with KK and suckout vs AA two times are you running good or bad? Are you playing / running well or poorly. According to an EV graph it will say you are running above EV despite being even, but when you calculate your expected rate of return with those hands based on the equity on opponents range you have an expectectation of almost +3 buy ins, but according to an EV graph you are even, and have an EV of -2 buy ins

So shouldn't this even out for everyone over the long run...i mean what you're saying is great for the short term, but in the long run these situations should even themselves out and EV should be fairly accurate, correct?

Variance is an incredibly difficult thing to figure out. It's also very difficult to take into account all the variables that can be attributed to running good or bad.

@Zep

1. In terms of hands dealt - how often you get 1010+, AK, etc I'd say those things even out for just about everybody.

2. In terms of how many situations you get where it's a free stack for you like AA vs. KK or Top Set vs. Middle Set - those things happen considerably more rarely. So it takes a much more massive number of hands than anyone will ever play for that to even out.

3. In terms of "All-In" situations where both players get their money in it varies. The frequency at which this happens in 6M, HU, and FR varies. It also varies based on the stakes being played. Lets just use an example that says Player A Get's all-in with an opponent once every 100 hands. That means 10 times for every thousand hands, 1000 times for every 100k hands, and 10k times for every 1M hands.

---

In particular in situation three it's really difficult to play enough hands in your life to get all-in enough times that luck will simply even out. Although that whole section doesn't even take into account the number of opportunities per 100 hands that you get an opportunity to make money with a big hand. It would be extremely difficult to calculate some sort of opportunities to make money index. In which every time you get a good hand bet and are called that would be counted as an opportunity. Opportunities on the turn/river would obviously be more highly valued because you make more money on those streets as the pot is bigger. All the math and calculations to figure all this stuff out would be enough to drive someone insane.

lol @ things evening out long term. the frequencies with aa, sets etc but the total amount of cash u get from those can never be the same for everyone. thats where luck comes in

i wish i had this nit's luck , just peeked into some of his hands the other day. super standard for this guy

https://www.pokertableratings.com/rep...mp;id=48611906652&hash=9078278784
https://www.pokertableratings.com/rep...mp;id=48610093595&hash=9077008461

bumhunting fag :/

Is there a possibillity to run it for like 1000 people and then overlay these graphs, because then you would see that majority of the people would be around the black line(the gaussian graph or whatever it is called in english)
Bigger sample just means bigger probability that you will run closer to your black line. So in theory after infinite number of hands everyone would be running as their black line commands.
Also this kinda relates to it http://en.wikipedia.org/wiki/Central_limit_theorem

Edit: Just now I realized that if you beat the limit you play on by 10bb/100, even if you run super super super bad, you still almost breakeven, so with rb its almost impossible to be a losing player.

Alot of people dont understand how much random variance poays into everything. When youre learning and try something for a few months that is bad but tun good some bad plays will be cemented in your mind. Also when you move up or take a gamble and run bad it can set you back a long ass time. You may never recover mentally.

Also these variance simulators illuminate how much luck plays a part even over what wed think is big samples. How many hands do we play over a month? They are statistically meaningless. Honestly, our poker playing careers are too short to really see the long term.

On August 26 2010 08:53 Maynard! wrote: Also these variance simulators illuminate how much luck plays a part even over what wed think is big samples. How many hands do we play over a month? They are statistically meaningless. Honestly, our poker playing careers are too short to really see the long term.

yup, the best we can do is try to play the percentages and pray.

i think what a lot of ppl dont realize though is that randomness plays a huge part in everything in our lives. theres variance in job interviews, etc. randomness affects everything in our lives.