ship or cawll ?

I would shove, and be very happy about it. he will call with any A and probably any JJ-KK and he might even call with some random T. 20 hands is enough for me. seems like a psycho.

looking at villain's preflop stats, i think this is more of a call than a shove. with 78% 3bet, villain is prob gonna have random 3x along with a bunch of total air. we might miss some value from Ax but we block that. i dont think villain will bet/call a shove with a value hand that loses to Ax, so we are hoping he has exactly Ax. also i would assume villain has more club-club hands than us.

18.0 AF, i kinda like a minraise here, since villain prob either has a hand thats not folding to any bet (Ax, clubs, 3x), or total air. so we can bluff some of our floats this way for a great price, though its hard to say how many floats we'll have.

20 hands is NOT a sample, you guys are complete fucking imbeciles, keep your shitty advice to yourself.

Do whatever you want, you have the best hand here more often than not.

20 hands is a very small sample, but if the statistics for 20 hand sample is extreme, you can get a statistically significant conclusion. For example, if you sit at a table never having played a single hand against an opponent, and they raise pre-flop every single time for the first 10 hands, then you can say that they have a pre-flop raise frequency of more than 50% at a significance level of 1% (ie. the probability that this person has a pre-flop raise frequency of less than 50% and just 'randomly' got 10 raising hands in a row is less than 1 %).

So even though 20 hands is a very small sample, a 75% PFR is extreme enough that we can still generate meaningful conclusions. Of course we can't conclude that his actual PFR is close to 75%, but we can be damn sure that he is not someone who has a true PFR of 20% or something.

If i were to print your comment the paper it is printed on would be worth more than what you said.

I'm assuming Cariadon is trolling? What you (HungarianGOD) say makes perfect sense, and we can use a binomial distribution to prove that over a 20 hand sample, if the stats are extreme then we can use them.

So over a 20 hand sample, and we assume that a 3bet level of 15% 4 handed is for a standard reg- thus we use a binomial distribution X~B(20,0.15) , and we calculate the chance that we get 15/20 successes (how he got 78% from 20 hands is weird though, also excuse the maths my knowledge is old haha)

using a calculator on the internet we get (this is greater than or equal too for some reason it copies only greater than) Cumulative Probability: P(X > 15)=3.18559001399876E-09- =0.0000000031855.......

so the chance of him having a true 3bet of 15% is roughly 1 in 313 million. given that at industry levels a chance of 1 in a thousand would be decent evidence used to prove that our original hypothesis isn't true, 1 in 313 million seems excessive. I literally quote HungarianGOD and say although we don't know his 3bet is really around 15%, we can be almost certain it is above 30-40%

On January 12 2015 17:47 the cleaner wrote: I would shove, and be very happy about it. he will call with any A and probably any JJ-KK and he might even call with some random T. 20 hands is enough for me. seems like a psycho.

I agree. Also, I would expect random fish to have a sizing tell here a decent % of the time (this looks like weak Ax)

So if a person has played only 1 hand and he 3bet the probability of him having a pfr of 100% is infinite ? There is a strong mathematical component to poker but it doesn't apply here. The guy is typing in bet amounts by the second decimal (27.91), he isn't fucking retarded or playing the game for the first time. He is just a drooling reg who sat down and started raising and 3betting because why not.

If you want to take advice from TheCleaner who has played 50nl for 10 years or from a HungarianDOG with no credentials then it is up to you. Personally, i wouldn't.

No need to be so hostile man! :-)

I am not talking about poker here, I'm talking about statistical significance and sampling, of which I would say you have a serious misunderstanding based on the first sentence of your last comment, (if you are trolling, you made me bite gj).

Statistics can't ever tell you "this person has a 3-bet % of 75%". What it tells you is that "the probability that this person has a 3-bet % of between 60%-86% is greater than 99%". The 99% here is the how powerful the statistical test is. From the exact same sample, you might also conclude that the probability is only greater than 95% that your opponents 3-bet% is between 70-79%. Because you have narrowed the range, your results are not as statistically significant (although in many fields, p > 95% is the cutoff for statistical significance).

The reason that you want a big sample is that you can get a more powerful statistical test. That is the only reason you want a bigger sample, there is no magic fairy dust that n = 1000 has that n = 30 doesn't.

Interestingly enough, even if someone 3 bets every single hand so has a 3-bet stat of 100%, no matter how big the sample size you cannot conclude that his 3-betting frequency is 100%. You will however be able to get an interval closer and closer to 100% that is statistically significant the more data you have

If you had a one hand sample and in that one hand sample someone 3 bet, you could say that the probability that his 3-bet frequency is less than 5% is greater than 95%, which maybe for you is not a strong enough statistical test. In fact, the ONLY interesting thing you can conclude with an extremely high level of statistical significance from a sample of 1 is that his 3-betting % is NOT 0. So though that seems obvious, to say that a sample size of 1 contains absolutely no info would be false.

You are correct that I have no poker credentials; I feel that most people still posting on this site are much better than me at poker. But the beauty of math and logic is that even though I have no credentials or reputation, I'm still right.

20 hands is not hard evidence, but its likely that he is a maniac

On January 12 2015 22:57 cariadon wrote: 20 hands is NOT a sample, you guys are complete fucking imbeciles, keep your shitty advice to yourself. Do whatever you want, you have the best hand here more often than not.

On January 13 2015 00:48 cariadon wrote: If i were to print your comment the paper it is printed on would be worth more than what you said.

and you called me a shitposter?

On January 13 2015 09:39 cariadon wrote: So if a person has played only 1 hand and he 3bet the probability of him having a pfr of 100% is infinite ?

...probability is bounded between 0 and 1...

On January 13 2015 09:39 cariadon wrote: There is a strong mathematical component to poker but it doesn't apply here. The guy is typing in bet amounts by the second decimal (27.91), he isn't fucking retarded or playing the game for the first time. He is just a drooling reg who sat down and started raising and 3betting because why not.

Player2 bets [$14.00 USD]

i believe math might be more important than your online tells

Hungarian or especially Noony I think you are operating under a misconception. if you have 20 hands on a guy and his 3bet is 78 you don't have a sample size of 20 for 3bet. 3bet is calculated only among hands that were possible to 3bet. in this case probably 7 3bets from 9 opportunities would round to that.

On January 13 2015 11:38 Santafairy wrote: Hungarian or especially Noony I think you are operating under a misconception. if you have 20 hands on a guy and his 3bet is 78 you don't have a sample size of 20 for 3bet. 3bet is calculated only among hands that were possible to 3bet. in this case probably 7 3bets from 9 opportunities would round to that.

You are correct in saying this, but I didn't mean to refer to the given example specifically. I was just trying to make a mathematical argument when a sample size of 20 could give useful statistically significant conclusions. I should have been more clear

The initial question was either to ship or call.

Your statistical, mathematical or whatever essay doesn't answer shit. The absurd pfr/3bet comes from the common practice of raising the shit out of the table because he can and because idiots make the wrong conclusions (oh wow this guy is a maniac psycho). It used to be, and apparently still is, common practice. Raising and calling are two different things, you can raise with any two cards but for calling you generally need a pair or better.

You can deduce that while this player is capable of raising light he is going to be smarter about his calls. So you can shove that smartass attitude where the sun don't shine and stop giving advice where it is not welcome.

On January 13 2015 11:38 Santafairy wrote: bla bla bla

Get your 10nl ass outta here, buddy.

sorry cariadon i guess things have changed now even 10nl players know how stupid it is to even for a moment consider saying something like "the probability is infinite"

You are right that the mathematical ranting tells us nothing about what we should do this hand; rather, it shows that you have no idea what you are talking about. People who are hostile/negative and people who constantly say irrational things are independently annoying; the combination is insufferable. If you would make your writing either more congenial or more intelligent, it would be really appreciated.

This used to be a forum where poker players convene to discuss the game of poker with the intention to improve. Now it has become a diary for some and dirty scrapbook for others - a bunch of has-beens and never-haves.

If you aren't playing the stakes (or worse, never have) then please keep your opinions and advice to yourself. This used to be the status quo.

"using a calculator on the internet we get (this is greater than or equal too for some reason it copies only greater than) Cumulative Probability: P(X > 15)=3.18559001399876E-09- =0.0000000031855....... "

The only way to respond this nonsense i know is to ridicule it, if you can't spot the obvious then you must be new to the internet.

you're doing gods work cariadon, keep itup

have to agree w cariadon
20 hands statistics are worth shit

On January 13 2015 09:39 cariadon wrote: So if a person has played only 1 hand and he 3bet the probability of him having a pfr of 100% is infinite ? There is a strong mathematical component to poker but it doesn't apply here. The guy is typing in bet amounts by the second decimal (27.91), he isn't fucking retarded or playing the game for the first time. He is just a drooling reg who sat down and started raising and 3betting because why not. If you want to take advice from TheCleaner who has played 50nl for 10 years or from a HungarianDOG with no credentials then it is up to you. Personally, i wouldn't.

go fuck yourself!

No hard feelings, i have nothing against you the cleaner. I just think you are stuck in the same "average reg mentality" For all i know you could be banking at 100nl.

its close, proly call

Yea probably call

not a bad try hungarian but starting with a prior belief of a normal 3b% then doing bayesian updating to get an estimate as more 3b come in is a better approach than taking the observed number as a point estimate around which to form confidence intervals.

I think there's surprising signal in 20 hands. vpip converges fast, and if you look at hh of how each hand was played there'll sometimes be something that stands out. preflop3b sizing in this hand is dodgy for instance.

hand is prob still a flat, but hungarian had closer to correct conclusions about sample size than cariadon in this case, imo.

instead of focusing on meager 20 hand sample stats i would focus on other aspects like his timing/betsizing etc for future hands (if villain is a likely reg) otherwise you are setting yourself up for extreme mistakes if you are making conclusions from 20 hands. What if this dood is a huge nit that happened to pick up 6 premium hands and 5 mid str hands inthat 20 hand span? then you are just punting into his monster range without even realizing it..I"d pay attention to what he shows up with rather than "78/60/x/x over x hands. x hands being super small #)

its close but just a call imo.

I count roughly 25 combos that beat you on this river (XXcc, AT, A8, TT, 88, AA, A3s)

Is there >25 combos of hands that bet/call and lose to AK?

AQ/AJ is 16 combos

A2s-A9s is 12 combos

So he'd pretty much have to get to the river this way and be bet/calling with 100% of his Ax combos for this to be a profitable shove

even if he's a maniac or in maniac-mode and has a bunch of A9o-A2o type hands, he'll still have to be bet/calling the majority of his Ax combos

would be a very ambitious shove imo


--

btw how do you guys factor the likelihood that this guy is a decent reg who is just on semi-tilt and maniac'ing it up (pretty likely given the extremeness of the sample) into your bayesian models?

btw how do you guys factor the likelihood that this guy is a decent reg who is just on semi-tilt and maniac'ing it up (pretty likely given the extremeness of the sample) into your bayesian models?

You could use a hierarchical model where on the first round it is determined, e.g. by a Bernoulli trial, whether he is a decent reg on semi-tilt. Then on the second round the the model takes that into account.

On January 13 2015 23:52 Smuft wrote: I count roughly 25 combos that beat you on this river (XXcc, AT, A8, TT, 88, AA, A3s) Is there >25 combos of hands that bet/call and lose to AK? AQ/AJ is 16 combos A2s-A9s is 12 combos So he'd pretty much have to get to the river this way and be bet/calling with 100% of his Ax combos for this to be a profitable shove even if he's a maniac or in maniac-mode and has a bunch of A9o-A2o type hands, he'll still have to be bet/calling the majority of his Ax combos would be a very ambitious shove imo -- btw how do you guys factor the likelihood that this guy is a decent reg who is just on semi-tilt and maniac'ing it up (pretty likely given the extremeness of the sample) into your bayesian models?

EDIT: ok my bad I didn't notice the stats. They are kind of relevant. Still its imo real easy flat.

pretty easy shove to me, the sizes makes really unlikely he has a best hand, plus stats and preflop size point to the fact he's an obvious fish and I really doubt that he will bet/fold with any pair getting 4 to 1 on the call,
even if he folds most of the time, the % he calls with random hands like 77, TJ or something will be > than 0 for sure and thats something you want to consider ( instead of finding hands that he will always fold/call )

it has to be flat almost always but i would at least think about shoving