BJ and statistics

[Flavio4321]

Every time someone says that online BJ is rigged, it is mostly because of a bad streak. After this person says ...." I lost 40 units in 300 hands ".... , someone in this forum tells this guy that losing 40 units in 300 hands is less than losing 2 SD so it is just bad luck. So a question came to my mind, how many hands do I need to lose to say that BJ is rigged at this "cheat" casino?. And I started to make numbers.
Let "cheat" casino have a BJ wich is supposed to have a house hedge of 0.49%. But the evil manager of "cheat" casino decides to make a little change on his software, so their BJ game has now a house edge of 1% (double profit for the casino), and SD of 1.2 like regular BJ. I deposit my money on this "cheat" casino and I play 10000 hands of BJ ( $1 dollar each hand for simplicity). I will lose normally something like (10000*0.01=100) 100 units. After losing I will post in this forum and someone like me will reply one more time.. You lose 100 units in 10000 hands, thats just 0.5 SD below spectations ( this smart guy is doing math with house edge 0.49% ). So I decide that it was just bad luck and I continue playing at this place that it is cheating and stealing money.
My point is, even with 10000 hands which is a lot for 1 person I think is very difficult to decide if a casino is rigged, So how do you decide where to play? where is the limit to the kind of answers .." just 2 SD below expectations"..?
Do you trust in the numbers that big software providers tells us in their websites?

(of course my math maybe wrong and dont judge my english, but I think my ponit remains valid)

 

[nafanny29]

I think it is very hard to detect cheating even with a large sample of hands if the casino is only adding 0.5% to their profit.

Im not convinced there are many casinos that deal truely random blackjack out there. Every different software seems to play differently and you get used to seeing the same sort of things on each type.

I cant figure out exactly why they "feel" non random but over many hands of blackjack online I can honestly say its not the same as real live play. I know its 10 times faster online hence many more hands/streaks etc but IMHO online blackjack is "controlled" in some way at most casinos.

Just my 2c

 

[caruso]

Do you trust in the numbers that big software providers tells us in their websites?

If you're mentally deficient, yes.

There is no way to prove anything unless the game is so heavily fixed that a relatively limited amount of data yields compelling figures, and this is quite a rarity. If you want a fair game, gamble at bricks & mortar casinos. You have no protection whatsoever online. Assume you are being cheated, because you probably are. In the olden days, Microgaming most certainly cheated at their blackjack game (there was MUCH player consensus on this), and I have compelling word of mouth evidence (uncorroboratable) that RTG were stone cold certified as cheating a few years back. Although my own results have been acceptable, I have seen such idiotic results from Crypto players as to be very convinced that the software can be manipulated. In my own unproveable opinion, Boss Media cheats beyond question and Playtech is not far behind. I have evidence (ie. proveable) that Gambling Federation cheats which is almost beyond question (5 SDs - Casino Bar was 6.8).

Considering the absurd risks involved, why anyone would gamble online for the sheer hell of it is an absolute mystery to me.

 

[Flavio4321]

lets do some scientific work

Both, caruso and nafanny29, have their reasons to believe that casinos are rigged. But non of them can PROVE or ar least give mathematical evidence of this. ( keep Gambling Federation aside ). I am looking for ideas on how to prove this, because if we cant show some evidence there are a lot of people that are going to continue gambling online and loosing money.

I dont like to say something that I cant prove, and as a science student It is my " way of life ". So I dont want to say this casino is rigged if cant put evidence on that.

( Of course I cant prove that casinos are honest either, but you know, everybody is innocent until they are proven guilty )

 

[caruso]

The best possible thing that could happen to the online gambling scene in terms of player protection against cheating software would be the development and implementation of the Fairdice project. Unfortunately, it doesn't seem to be something which is going to happen any time soon since it's a part-time venture for the folk running it. I've tried to get an idea of the stage they're at currently, but communication with them has been patchy to say the least.

There was discussion at WOL:

You do not have permission to view link Log in or register now.

Their website:

You do not have permission to view link Log in or register now.

Until they get up and running (if they ever do), one is left with the PriceWaterhouseCooper "software audits" - and I cannot think of any better reason than that to push for the Fairdice project's successful implementation and acceptance. It would be an enormous move in the players' favour.

 

[Freudian]

You don't need a huge sample size to show something is off, since the results have to be so extreme to show a deviation of 3+ SD (which is when you should start paying attention) if you have a smaller sample size. All a statistical analysis tells us is how likely something is. We ourselves have to make up our minds on how to interpret an unlikely event.

And of course I don't trust the numbers the operators tell me. I may be more inclined to trust the numbers of a specific operator based on my own and others experience with them, of course. But even the biggest scum among operators will tell you numbers that look fine.

 

[nafanny29]

Another discussion on the riggedness or not of online gambling. Im a firm believer of the fact that "nothing online is exactly as it seems" but I frequently gamble on the net!!

I am with Caruso on the Fairdice project, or any conclusive study/proof that online casinos/software are rigged or not.

I fear that when a detailed analysis is made of the major players/software platforms, it will result in a unanimous "yes its rigged" verdict.

On a UK side note, once online gaming is regulated and allowed in the UK no casino registered here will be able to get away with cheating thats for sure, the one thing the Brit government are VERY good at is regulation and control.

 

[Clayman]

I dont like to say something that I cant prove, ... So I dont want to say this casino is rigged if cant put evidence on that.

Me neither.

When I first started playing a million dollars ago I had no idea what to expect. I knew nothing, absolutely nothing, about probability, standard deviations, etc. I barely understood house advantage.

I'm so paranoid, I recorded the actual cards dealt because I figured they could doctor the logs with different cards.
I was so paranoid I considered video taping the monitor because no one would believe me if what I said differed from the logs.
I'm still so paranoid if I don't see all 8 decks spread out at a table before I play I'm likely to think a couple of 5's have fallen on the floor.
After a while I realized the logs always agreed with what I said and they at least weren't changing the logs on me. So I just stuck with recording the results of hands since I had to know how much I wagered anyway. It was the only defense I could figure for when they would cheat me. And, believe me, I knew they would. I figured in this way at least I could prove they cheated me when they did.

So I just kept writing down my W's & L's and, as the column grew longer and longer and I picked a little up here and there off the Internet, a funny thing happened. I slowly noticed, much to my surprise, things were pretty darn close to how they were suppposed to be.

This, while nice, proved nothing to me. Even so, I could be just lucky or, if by some chance they weren't cheating, they could start at any time and I wanted to know when they did. So I continued writing my W's & L's. And my column grew longer and longer.

Eventually I got more sophisticated and made separate columns for each different software provider and for the different BJ games within each. This way I figured I'd know better exactly which provider was cheating me and in which game. Unfortunately, even then, every game came out pretty close to where it should. Sure I'm over 3 standard deviations to the good in Micro single-deck but I'm over 2 SD down in Crypto multi-deck. It still proves nothing but it's all I have.

The only reason I record a hand at all is so I can catch them cheating me when they do. I don't really care what Price Waterhouse or True Gambler or ecogra says, what the casino says, or, for the most part, what other people say, though hopefully I listen to all of them with an open mind. Ultimately, it's an act of faith what you choose to believe but a little evidence can't hurt.

So my column continues to grow one hand at at a time, and, while I will never know whether someone else is ever cheated or not, I hope at least I will know when I am.

 

[jetset]

QUOTE I listen to all of them with an open mind UNQUOTE

Clayman, this is one of the reasons why I always read your posts with close attention and respect - they are invariably balanced, reasoned and based on experience and recorded fact over time, rather than unproved but confidently voiced assertions and *gut feels*.

The concept of effective, independent and ongoing monitoring and analysis of gambling softwares, preferably through an unbiased, properly administered, practical and thoroughly tested player input system has real value, if only to reassure players baffled by the often conflicting opinions of posters. There is obvious player appeal in showing that particular softwares meet acceptable standards of fairness over time in a system that obviates any perceived conflict of interest i.r.o. operators or software suppliers.

However, it does not appear that this is going to evolve just yet. We have seen the TG system languish after some pretty alarming faux pas and underfunding, and the Fairdice system (which I introduced as a topic on WOL I believe) does not seem to have advanced much beyond the realm of enthusiastic amateurs in development, verification or implementation.

Bethug of the OPU has now come forward with a presumably contracted and qualified expert using undisclosed methods to carry out a technical study that will eventually be presented by Bethug, allegedly to obviate operators "buying it out" (and btw I have yet to see any evidence whatsoever that industry people have *bought out* testing initiatives to keep them off the market, attractive as that idea might be to those who want to believe that online gambling is one huge scam.)

I think I will continue for now to rely on the opinions of recognised and trusted experts who back conclusions with detailed evidence, and the results of professional analysis of massive amounts of data on every single transaction to which PricewaterhouseCoopers put their not inconsequential name and international reputation.

 

[Vesuvio]

I must admit all inclination I had to believe that Bethug, however misguided, really had the player's best interests at heart disappeared when I saw the new website for his 'system' which asks for credit card details & charges a 'processing' fee. It also seems to have some sort of guarantee system which will encourage players who've been successful to donate part of their winnings - a common scam with progressive systems as players will more often that not win money in the short term. Unfortunately giving money to someone else just increases the long term losses.

Getting back to the fairness of on-line BJ - except for one or two providers I think you do get long term results in the expected range, but I don't trust any of them to deal random cards each hand. It must be possible to somehow test the true randomness of RNGs - it always annoyed me about the PWC reports that all they looked at was the correct results long term and over a 'reasonable' shorter period of time. Couldn't they hire a computer programmer (or statistical analyst) to look at the actual functioning of the RNG?

What we really need is for someone in the know to spill the beans on exactly how casino software functions - something which the majority of casino owners probably don't have a clue about. Perhaps the closest we got was when one group of casinos defected from Microgaming claiming concerns over the RNG. They did defect to Playtech, though

 

[nafanny29]

Getting back to the fairness of on-line BJ - except for one or two providers I think you do get long term results in the expected range, but I don't trust any of them to deal random cards each hand. It must be possible to somehow test the true randomness of RNGs - it always annoyed me about the PWC reports that all they looked at was the correct results long term and over a 'reasonable' shorter period of time. Couldn't they hire a computer programmer (or statistical analyst) to look at the actual functioning of the RNG?

Very good point Vesuvio. Ive seen loads of posts about BJ turning sour when the bet size is increased. Maybe the "RNG" deals better cards to the smaller wagers and deals poorer cards to the big bets. Overall the statistics hold up under scrutiny etc. Just a thought

 

[Flavio4321]

Maybe the "RNG" deals better cards to the smaller wagers and deals poorer cards to the big bets. Overall the statistics hold up under scrutiny etc. Just a thought

If you read the PWC payout & RNG report at any MG casino or at CON, you will see that they audit average payout (being the percentage of winnings to wagered amounts). So, If the software deals poorer cards on big bets we woud have seen it in this reports. (if we belive in PWC audit).

Bethug, if you have an idea on how to test casino BJ please tell us, but saying that you have paid this math wiz to verify games is NOT the point of this thread.

This thread intention when I started it, was to look for mathematical methods or any other ways of verifyng casino BJ, if you fellow players have somthing to say about this say it, but avoid saying this or that software is rigged if you dont have proof of it.

Bryan and Bethug, if you want to have an argument between you, please start another thread. Bryan I know this is your forum so you do whatever you want, but this thread wasnt about onlineplayersunion legitimacy. So It will be a good idea to start a new thread about that, I will be more than happy to post on that thread.

 

[bethug]

sorry flav

 

[eek]

The best potential system is the fairdice system, since it works in realtime and monitors each round.
It seems to be along the lines of my comp sends them an encrypted number, and they send me one.
These two numbers combine to create the RNG result.

I'm not sure how the seeding/calculation for the final RNG result works tho, and who sends who 'their' number first? Since whoever has both numbers first (and has not yet transmitted their own result) has a potential advantage in the timing difference.

I presume this is the bit where the developer is looking for an acceptable standard of security.


Other systems like truegambler or monitoring results on a spreadsheet are more retroactive, whereas Fairdice has a more proactive potential.

Spreadsheet macros are useful for analysing a session Flavio, but you would have to learn macro programming.(which is not very hard btw)
Macros are hugely versatile.
The Wiz gives useful stuff on probability, studying his casinobar project will give you useful pointers on how to spot non-random deals.
(This project btw was for only ONE type of cheat only-dealing seconds)

You do not have permission to view link Log in or register now.

 

[bethug]

the wiz didnt spot nothing wrong with the software until, someone from the inside told him i believe.

Its not just blackjack, its baccart and other games that need to be tested and if fairdice can get up. That will be great

but how many software companies going to let fairdice in there so called random genertors

Douglas Ray of Fairdice says: "This will deal a bitter blow to conmen and criminals in the gambling industry, and clear the way for the honest online casinos who give their players a fair chance. With the launch of this initial implementation of the Fairdice cryptographic protocol," he continued, "Project Fairdice is offering casinos another way. We're saying to all the casinos out there, 'Come get involved in the project. Enable your websites to use the Fairdice Protocol. If you are not rigging your games, you have nothing to lose. This is your chance to really prove to your players that your random number generators are not fixed."
Additional information about this new software security system can be found at

You do not have permission to view link Log in or register now.

 

[Clayman]

Ive seen loads of posts about BJ turning sour when the bet size is increased. Maybe the "RNG" deals better cards to the smaller wagers and deals poorer cards to the big bets. Overall the statistics hold up under scrutiny etc.

I actually have higher win %'s and lower losing %'s at higher bet sizes ($20-$100) than at lower bet sizes (<$20).

In a 199 $100 bets (my highest bet) I have 98 Wins including 9 BJ's (49%) and 81 losses.

(The above only for Micro SD).

 

[Clayman]

This thread intention when I started it, was to look for mathematical methods or any other ways of verifyng casino BJ, if you fellow players have somthing to say about this say it

The simplest way I know requires no high mathematical skills and no spreadsheet skills.

All that is necessary is one keep track of the number of winning BJ's and winning, losing and tieing hands for all non-doubled and non-split hands. I say winning BJ's because tied BJ's would count as a tie.

Do the same thing for doubled and split hands separately. Obviously you will not need a BJ column here.

I don't care if you make hash marks on a sheet of paper, flip cards over or how you do it. I happen to write BJ, W, L, T down in a column and count up the number of each later using a "=DCOUNT" formula to do it.

You do not even need to record the amount you bet if you don't want to.

If you have this, you have everything. If you do not, you have nothing. Everything else is bells and whistles.

 

[caruso]

The best potential system is the fairdice system, since it works in realtime and monitors each round.
It seems to be along the lines of my comp sends them an encrypted number, and they send me one.
These two numbers combine to create the RNG result.

Eek, do you know anything current about the state of the project? I've drawn close to a blank and fear it's going to end up on being put on permanent hold or at best be years in the making. Several months down the line and it still seems very embronic, from the little I've been able to ascertain for myself.

 

[Flavio4321]

All that is necessary is one keep track of the number of winning BJ's and winning, losing and tieing hands for all non-doubled and non-split hands. I say winning BJ's because tied BJ's would count as a tie.

Why do you separate non-dubled and non split hands?

Last month I started to keep records, but I simply registered number of initial bets, units lost/won. I did this because I figured that after playing a couple thousand hands I could calculate the "house edge" for my game in a simple way
HE=(units lost / inital bets) * 100 %

Can you tell me calyman what do you think about my records and why do you think that it is necesary to keep doubled and splited hands records aside?

 

[cipher]

Hi Clayman:

Quote: "You do not even need to record the amount you bet if you don't want to."

With a great many of the casinos on the Internet I've found that there appears to be a direct correlation between the amount wagered and the probablity of a player winning on a higher wagered hand versus a lower wagered hand. By not recording the amount wagered your data would be incomplete at best. Have a good one.

 

[Clayman]

I figured that after playing a couple thousand hands I could calculate the "house edge" for my game in a simple way
HE=(units lost / inital bets) * 100 %

I just keep track of the doubles because I like to know the payback assuming I bet $1/hand so, for example, in 100 hands I've probably wagered about $110 because 10% of my hands are doubled or split.

But as long as you mean by "initial bets" the number of initial bets you have made (in other words the number of dealer upcards you have faced, i.e. the number of hands you have played) rather than the total dollars wagered on your initial bets, you are fine. Even simpler, really, than what I said although I would find it more difficult to keep a running total of net units won/lost while I played. (I'm assuming by "units lost" you really mean "net units won/lost".) Just curious, how would you know if you had mis-added somewhere along the way? Any checks and balances in your way of doing things?

So if I play 100 hands and I have 5 winning BJ's, 39 regular wins, 48 losses and 8 ties and 10 of those 100 hands were doubled of which I won 6 lost 3 and tied 1, I would say I won 7.5 units + 39 units + 6 units minus 48 units minus 3 units for a net gain of 1.5 units. And it would not matter if everyone of those hands was a different amount bet. And I would say my payback is $1.5/$110 wagered no matter how much I actually wagered or how much I actually won or lost.

But you're right - all that 1 in 20 crap is based on net units won or lost divided by number of hands played.

Hope this helps.

 

[Flavio4321]

I just keep track of the doubles because I like to know the payback assuming I bet $1/hand so, for example, in 100 hands I've probably wagered about $110 because 10% of my hands are doubled or split.

From wizard of odds ...."The house edge is defined as the ratio of the average loss to the initial bet. The house edge is not the ratio of money lost to total money wagered. In some games the beginning wager is not necessarily the ending wager."

So, if you want to know your "payback" it is ok the way you are doing it. But I want to compare my results with theoretical calculations, the only therotical number that I know for BJ is the "House Edge". That is why I only kept count of "initial bets"

But as long as you mean by "initial bets" the number of initial bets you have made (in other words the number of dealer upcards you have faced, i.e. the number of hands you have played)

yes, thats what I mean for initial bets

(I'm assuming by "units lost" you really mean "net units won/lost".) Just curious, how would you know if you had mis-added somewhere along the way? Any checks and balances in your way of doing things?

Yes, I mean net units lost/won. It is easy to kept units count for me because I flat bet every session, the only thing I have to do is divide my net total result by my bet to get my net units total. If I decide to change my bet size I simply write down my balance at that time.


Anyway you have more information than me in your logs, but I am to lazy to do all the stuff you do.

Cipher, some time ago you emailed me one of your log files, I didnt find any correlation between size bet and cards dealt, but of course I am not an expert so maybe I couldnt see this correlation. One more thing, when do you say correlation do you refer to the statistical definition of correlation, because I remeber from my statistics course how to make a correlation test, I will be more than happy to make this correlation test on your logs, but you have to tell what to look for. For example you can say that when bet size is increased, dealer gets more 5s or somthing like that.

 

[Clayman]

With a great many of the casinos on the Internet I've found that there appears to be a direct correlation between the amount wagered and the probablity of a player winning on a higher wagered hand versus a lower wagered hand. By not recording the amount wagered your data would be incomplete at best. Have a good one.

Hi Cypher
Most ways of statistically analyzing BJ assume flat-betting, as I'm sure you know. Underlying that assumption is that the cards are dealt randomly and do not know the amount bet.
So one can believe what one wants to. If one believes what you are saying might be true then of course you would want to record your bet amounts and the record of same.
If I owned a casino, I would release the BJ, W, L, T record by amount bet. I don't know why they don't.
Failing that, as you can see from a few posts above, I do keep track of wins and losses by bet size. And I have no significant variance at any bet size with a reasonable number of hands. In fact, personally, I've been luckier with bigger bets than smaller ones.
I guess you are saying your system can not only identify when to make "bigger bets" you will win, it can do so in the face of a "direct correlation" to the contrary. (I'm assuming you mean the player has a lower, rather than higher, probability of winning with bigger bets.)
And of course I'd be really interested in seeing any evidence of this correlation, especially since every one of the strands I have analyzed that you have posted would tend to show just the opposite. After all, you don't win a session by losing your biggest bets.

 

[Clayman]

Anyway you have more information than me in your logs, but I am to lazy to do all the stuff you do.

Sounds like you are anything but lazy and you are doing just fine. Look forward to hearing some results in a few thousand hands or so!

 

[cipher]

Hi Clayman:

Quote: "especially since every one of the strands I have analyzed that you have posted would tend to show just the opposite"

That's what happens when you become proficient at reading trend analysis and you're able to recognize the seeding and dumping. But for the gamblers that are not familiar with these techniques (99.99%) they're pretty much dead meat on a hook. Have a good one.