PFF. lol

Josem, I see that you have now answered my question about whether or not your heartlessness knows no bounds.

Interesting...

Hello!

I am a long time professional limit holdem player. I have won a lot in my career and I certainly suffered some tough bad beats.

Anyway, what I have experienced in last few months is absolutely unprecedented. So I decided, since I have absolutely nothing better to do, that I will try to prove my case. I will be extremely happy if I am unable to prove it.

Some questions and observations:

1. How many hands do I need to make it statistically significant? I have 13k in my current database (limits 10/20 and 20/40, mostly shorthanded), but this is roughly from this month only. I have quite a bit more where the results was more or less just as bad.

2. Could PokerTracker help (or deny) my case? I am not an expert user of it so I am asking anyone more into it.

Some basic stats about these 13,005 hands:

Hands won: 2,413 (18,55%)
Amount won: -6,946.15
BB/100: -2.18

VP$IP: 29.90
PFR: 20.65
WTSD: 40.58
W$SD: 52.26

All seem more or less normal, except of course for the winning rate.

3. I am also a programmer, so if PokerTracker can't do it, I will go through hands by my own. I was thinking along these lines (yes, for the time being it is very simple):

- Take just the hands that went to showdown.
- Determine equity preflop, on the flop, on the turn (btw, is there any "equity analyzer function like PokerStove that can be called within software?)
- Determine the winner(s)
- Compare preflop (flop, turn) equity with actual results
- Calculate, how statistically significant these results actually are

4. Some general observations of rigging on this site:

- Preflop is from my experience normal. There is no point in analyzing preflop.
- Flop is mostly normal, but somehow it fits me more when my cards are already in the muck (I know, it could just be a feeling, but how can I analyze that?)
- The real horror is on turn and river

I will be very glad for constructive comments and any suggestions you may have.

Thanks.

Yeah, I'm a pretty open sort of guy.

Your main problem is accounting for card selection.

How do you come up with a 'correct' figure for a particular hand winning unless you know exactly which hands other players are going to run with.

I would suggest that your best bet is to go through your hand history and, for any hand you find problematic, go through and see if the proportion of times it wins is the same as the proportion of times it would win against all hands that went to showdown.

You could similarly test your hands against bought in HH's from similar game/stakes at other sites.

What you need is some kind of Real Deal for online play. Has anybody thought of that?

Real Deal?

That sounds like an interesting and promising idea.

Please tell us more.

1. Yes, you are right, people who draw on flop and turn will fold the river when unimproved, so when the showdown actually occurs it will show tight player being unlucky more than his fair share in any case.

I realise that. On the other hand, I suffered so many sick beatings, that I am quite sure that my equity change from preflop to river is going to be absolutely dramatic.

2. How should I accomplish that? Manually? I can't see it doing with computer. Maybe I am lazy, but I am not prepared to do it manually.

Essentially the same 7-8 guys constantly defending OP and new guys coming practically every day with doubts. Amazing

It's completely natural for somebody who is losing to try to find reasons why, especially if they have played an inadequate sample size, which is incredbily common. There's a reason why there are more established posters who believe the games are legit and new posters who think it's rigged, primarily due to the experience. Personally, I've played about a million hands of poker. I can't speak for anyone else but I think differences in experience have a lot to do with it.

If you really take a look at the reasons for people's doubts it's not all that amazing. Poker can be a very frustrating game. I've taken numerous breaks from it because of the ups and downs, but still maintain overall profitability. I simply do not believe the games are rigged because I have a math background and am well aware of what is possible when you really look at the statistics. Very common for people to notice the times they take a bad beat, but when they get it in as a 90/10 favorite and win the majority of the time, they don't think twice.

jagnje I'm sure I speak for many people when I say it's refreshing to see somebody who is suspicious actually show willingness to do the research instead of just concluding that the games are unfair. I have little doubt that you have probably run badly, but with only 13k hands my biggest concern is that there just isn't an adequate sample size to be able to test anything in great detail.

It's a lot easier to analyze hands that are all in before the river since that's the only situation in which odds can be calculated and a hand is guaranteed to show down. But since you play limit obviously that's rarely ever the case. I'm only familiar with analyzing no limit, where EV calculations are a lot easier to do.

There are certain tests you can do of course, seeing how often you hit flops in certain situations compared to expectation, but like you said your biggest concern is turn and river and obviously in those spots you're usually only going to see your opponents cards when drawing when they actually hit.

Another thing that jumped out at me, unrelated to the statistical analysis, was that you have a vpip of approximately 30. It may be different in limit, and it's possible that you're playing shorthanded, but that number seems very high to me and could be part of the reason for the struggles. I've found in no limit at least that you can pretty easily get away with a 30 vpip style playing live but the online games are much tighter and simply don't allow for profitability that way. They haven't since approximately 2005 anyway when the TAG style became much more common.

Anyway I'm obviously rambling at this point but I encourage anybody to throw ideas out there to help jagne with his analysis.

FYP

Yea just like Internet Poker you never know who or what your opponent is.

+1

So your talking about yourself here. Rather than believe the obvious that the sites are rigged you'd rather search for any reason other than that because you don't want to believe the horrible truth.

I'll repost something, mutatis mutandis, that I posted earlier in the thread and which you, as usual, ignored.

----------------------------------------------------------------------------

So, let's just firm up out positions here.

You and your fellow rigtards believe that all the major sites are certainly rigged despite there not being a scrap of credible evidence to support that notion.

I and the other anti-rigtards believe that the major sites are very probably not rigged because there is not a scrap of credible evidence to support the notion that they are.

Can you see why normal, sane, intelligent, people without your gifts for seeing strange patterns that defy mathematical analysis are going to be rather happier with our position than with yours?

----------------------------------------------------------------------------

This applies to all the new accounts, whether genuine or gimmick, which keep popping up with the same tired old story: "I'm losing when I should be winning and I think that the site may be doing something odd".

All you need to know is whether or not it can play better poker than you.

If it can; stop playing against it, if it can't; continue.

It's hardly rocket science.

Actually the VAST majority of people that I speak to about Online Poker are saying that it's rigged and they are sane intelligent people.

btw I do not claim any special gifts other than common sense given all the bs I've seen playing Online Poker.

Lies. Sane, intelligent people tend to hang around with sane, intelligent people.

And I defy your claim of having common sense.

OK, let's forget everyone else for a minute and also ignore the genesis of your concerns.

Just look at the two senetnces below:

1) You believe that all the major sites are certainly rigged despite there not being no credible evidence to support that notion.

2) I believe that the major sites are very probably not rigged because there is no credible evidence to support the notion that they are.


Contrasting those two statements, does your position not give you even the slightest cause for concern?

Yes, I know that NL analysis is much simpler. That's the precise reason why I am more suspicious of sites rigging limit than NL. Limit holdem is still a goldmine - or should have been - despite the decrease of popularity

Like I said, 13k is only a sample to test any algorithm on. The games are 95%+ shorthanded, so 30 VPIP is not too loose (maybe it is even too tight).

I looked into previous database (before May 2009, limits 2/4 to 5/10). The result is apalling: 135k hands, -8,5k.

Yes, I know, these limits are rake very unfriendly, but still... Put together, I have around 162k hands, since there are still 14k 5/10 hands from May and June 2009. Is this enough to base my analysis on?

Yes, I could test for specific flops. I suspect that aces fly on the board when I don't have them, I could probably test that with some degree of certainty. Another idea: test for Ax vs Ax situations when A on the board and being headsup. This situation is quite unfoldable in limit poker. I could also test for differences between limits up to 5/10 and higher. I had substantially worse luck in higher games.

Problem is, if they rig it, they rig it very smartly, so that it is hard to prove anything. FWIW, I was in the "they can't be rigging it" camp for over 5 years and I wouldn't have lost so much if I actually believed that they are capable of doing it. Of course it is laughable when someone posts a bad beat hand and says that it is rigged. It is equally laughable, though, that one should gather millions of hands of bad beats to prove that the site is rigged, since nobody could even stand so much bad luck. I managed to gather so many hands because opponents really play dreadfully (and well, I also never had to deposit anything after initial deposit).

FWIW: With rakeback and bonuses I am actually up in this period, so this is not from a losing player (overall). I just feel that I am not getting my fair share and it seems stupid to continue to play without even trying to analyse what is going on.

If anyone has any PokerTracker query that could help me discovering anything interesting, please let me know.

I hate to say this as you are clearly not one of the usual breed of rigtards but I have to ask:

Are you sure that your 'bad luck' is not down to the competition getting better?

It seems to me that the game continues to improve as people study more and I pretty much believe that in games such as this where the instantaneous result is as much down to luck as skill, bad luck is largely indistinguisable from less than optimal play.

I think we are about to enter an era when being a nit will not be enough to enable players to continue to make a profit. As this happens we will see more and more players who have been profitable for, maybe, years starting to turn into marginal or even losing prospects. Without a doubt some of them are going to become convinced that it is the sites that are causing their problems.

As you suggest it is hard to get an adequate sample size to prove that something is amiss simply because there is so much poker being played that the normal effects of varience will mean that someone will suffer the most appaling bad luck without anything untoward going on.

So unless you can show something so outrageous that it could not be expected to happen in a trillion hands you are still going to be told: varience.

That is why those of us who support the view that online poker is probably not rigged rely just as much on the logic that says that the risk to a site of being found out is too great to be outweighed by a short term increase in profit.

I commend you for wanting to do research. You can do things with 13K hands but you really need more to get good results. With only 13K hands, your significance levels are going to be very large (allowing high percentage variance from expectation) and probably won't tell you much.

You can do a board card analysis for all flops, turns, and rivers that were seen. They don't need to go to showdown. If your interest is in showing the statistical validity of the deal, do this without regard for your hole cards or anyone else's hole cards, just analyse the board. Since you don't know all the hole cards, you need to treat them all as unknown. You would do a completely separate analysis of your hole cards wins vs. expectations but you can only do that when the hand is all-in preflop with a single caller (you can be on either side of it). Anything other than that and the analysis is full of card removal effects due to player decisions and you can't evaluate the deal at all.

There are also some card removal effects that come into play when analysing the deal of the board cards since players choose to see the flop and choose when to continue, but the effects are small in large samples. In your sample they could be more pronounced depending on the particular game style your sample comes from. For example, it is known that in all variants of holdem low cards will hit the board very slightly more than high cards since players like to see flops when they hold high cards. But with a full ring, this effect is very small (less than half a percent difference between Aces and dueces).

Here's an example of how you might analyse the board. This is some work in progress of mine so I've omitted much of it, and ignore the numbers as this is just a test run on a fairly small sample (~2 million boards on one site):

Then you do similar things on the turn, looking at actual vs. expected for things like:

This should give you some ideas. I'll be publishing a large study using this and other analysis on large samples from various poker sites, soon. It's impossible for a site to manipulate the community cards without the skew showing up in tests like those shown above. In the above you'll notice the known card removal effect I mentioned when you look at the frequency of AK2 boards, since high cards flop less than low cards. You have to take these into account, and I've discovered other such effects too, that are consistent with a fair deal.

For the rest of you, the above is just a teaser of things I've been promising.

I was going to reply to this, but there's no point wasting the effort so I'll just say

"lolololololololololol"

I don't have anything for PT, but if you can download the HeM trial and import your hands, I am working on something that might be what you are looking for.



This particular part of the program collects the hand information from all hands that made it to the river and the hole cards for at least two players are known. From there, it calculates the equity % and compares it to the actual win %(user can decide if the equity percent is calculated pre-flop, on the flop, or on the turn). All equity calcs are derived from heads-up matches, so, for example, if three sets of hole cards are known, the program will create 3 heads-up matches, with the third set of hole cards in each match up used as dead cards.

Sadly, generating this graph takes a significant amount of time--especially for pre-flop comparisons. The graph above took my 2.0GHz, dual-core Athlon about 18 hours to run. I'm trying to get it multi-threaded to improve performance, but I'm just not that familiar with multi-threaded programming.

Anyway, I will be making this program available in the near future, along with the source code (don't want anyone saying that I'm rigging things too). It might at least be a starting point for your investigation.

-red

**edit** Forgot to mention that even though the title bar says it is for Full Tilt, the software will be able to run calculations for any site that is supported by HeM. The title bar is actually changed, depending on which site the user is running the analysis on.