OK, I'm back earlier than expected. Tach and others, thank you all for the interesting contributions. Tach, to Fedis48, in post #79, you said Let's not, since you are clearly unable to understand what I am saying I think all people here, scientist or non scientist, do not understand what you are saying. We are all agreed on the fact that there was NO mechanical bias - it was a fair roulette wheel. Therefore, it's results were always random, along the probabilistic odds of said roulette wheel. The big issue here, is that you are positing that the algorithm was somehow able to process these results, and give meaningful information about entirely random, future events. Please don't miss what I just said. You MUST surely agree that to give continuous losing results, i.e., red for actual black, odd for actual even, etc, off trend, and for 30 runs in a row (or any runs in a row), is as valuable a tool (it was to you, as you claim) as to give winning information. Even if you don't agree with this, lets just stick to your claim that it gave repeated LOSING information, i.e., continually picked the loser. What nobody seems to understand, is, HOW was it able to do this ? Surely you can see you are implying some sort of precognition on the part of the algorithm - a precognitive ability to lose rather than win (which I maintains is the SAME as to win, but no matter) How was it able to do this ? What force, what influence, what information flowed backwards / forwards in TIME AND SPACE, as surely it would have to, for this to occur ?
You claim it is a physical problem and not a mathematical one, yet you also agree that there is no physical interaction between the ball/wheel and the algorithm. So, as has been asked by another: how does the algorithm in any way affect the 50:50 odds of the RED/BLACK result? Do you at least concur that the result is 50:50 as to whether it lands on Black or Red? And, as I have explained, the only way you can introduce non-randomness into the actual result is through a physical interaction, which you have agreed did not happen. So you are left trying to explain how this algorithm affected the result of the ball/wheel? And the point is that the algorithm was irrelevant - as the result is always 50:50 red/black (ignoring zeros). Any perceived success or failure is just that... a perception. In the long-run the algorithm would have generated success of 50%. No, because his was based on the mechanical bias of the wheel, where the odds of red/black are NOT 50:50. But you have excluded mechanical bias. You have said that the algorithm in no way acted physically on the wheel/ball. Presumably you are agreeing that the actuall ball would land on 50/50 Red/Black? So, despite all the maths and explanations to the contrary, your only support seems to be anecdotal evidence from yourself that you won 30 times in a row, that flies contrary to math and to physics. This is a clear implication of what you writing... even if you, as it seems, does not understand it to be.
This sounds extraordinarily implausible. And you said you did this for 30 runs straight. Another contributor rightly put the odds of this at 2.44 billion. Can you please specifically address the following; 1) Prior to you starting your betting run of 30 (in your hour) for how many spins did you observe the wheel doing the same thing ? 2) You had certitude (as you stated in another post) .. yet you only won 800 eu ? What were you betting with, cents ? What was the house limit ? Anybody in their right mind, if they had the certitude that you claim to have had, for 30 runs in a row, would quickly and inexorably escelate therir bet to the house limit. Assuming that was 1000 eu, (on red/black, etc) one should have won somewhere in the order of 25,000 - 29,000 eu. Please provide clear answers tto the above. Thanks.
Yes, I now remember it, we talked about the fact that you are clearly not comfortable with experimental physics. But the "software" is a simulation of the physics of the roulette. Though the wheel bias was removed after the fiasco with dr. Jarecki breaking the bank repeatedly, the bias cannot be eliminated, the same way you can NEVER eliminate systematic errors from ANY experiment. The people who wrote the simulation must have tried to exploit such a BIAS, be it the depth of the shoe, the roundness of the ball, the way the croupier throws the ball, etc. So, their program is very similar to Jarecki's except it is not based on the wheel and, more importantly, because it is incorrect (while Jarecki's was correct). So, like in any experiment, the systematic errors are there, the cannot be eliminated but, in this case, the could be exploited, PROVIDED that their simulation were correct. Since it isn't correct (I proved that through experiment), it means that their probability of success doesn't start at \(\frac{1-P_0}{2}\) as in your idealistic approach but at \(\frac{1-P_0-B}{2}\) in a realistic case. The reduction by the bias accounts for the fact that their algorithm is wrong. In the case of dr. Jarecki, his simulation gave him a slight advantage, \(\frac{1-P_0+B}{2}\), good enough to break the bank in multiple casinos. See above, the WHEEL bias has been eliminated (well, more correctly , greatly reduced), the other sources of bias have NOT been removed, they can never be eliminated. If you were an experimental physicists, you would have known that systematic errors can NEVER be totally eliminated.
No, I have excluded the wheel bias (more correctly, the casinos minimized it after the experience with dr. Jarecki's simulation), you can never remove all sources of systematic error from an experiment. Read my answer to Fednis48.
You are not paying attention, we agree that there is no wheel bias. Yep, "wheel". ONLY. There are a lot more parts to the experiment. Read the answer to Fednis48.
red/black and odd/even interspersed with a lot of other different bids that I did not participate in? Did you read the part where I told you that I stayed an undetermined number of rounds in-between? Your calculation is seriously off. Each draw is independent of the previous one, so why are you multiplying them? This is not a Markov chain, at least not the way you are formulating it.
Ok. So your position is that even with the wheel bias fixed, there are other forms of bias that can be exploited. That's good to know - it's not at all what I thought you were saying. In terms of the equations I wrote, would it be fair to say that you disagree with the assertion that each spin is an independent, random event with \(P_{red}=P_{black}\)? If that's not what you're saying, please explain what you mean by "the bias cannot be eliminated" in terms of quantities more basic than \(P_{yourcolor}\). I spend a lot of time thinking about ways for my experimental colleagues to suppress systematic errors; I know they can't be totally eliminated. I just thought (wrongly) that we both agreed the biases in roulette were small enough to ignore for all practical purposes. What puzzles me, though, is that you say there is no bias in the wheel, but there is still bias in roulette. If it's not the wheel that's biased, what else is there in the experiment? The algorithm is just a piece of software that tries to guess the outcome of the experiment, so algorithm bias doesn't count for this purpose. Unless you want to talk about the ball and wheel as two distinct systems (which you're welcome to do, although that would seem a little pedantic to me), roulette IS a wheel, and minimal wheel bias means minimal bias of any kind. Like you say, each draw is independent of the previous ones. This means it doesn't matter how many rounds you sat out, or how many types of bids you weren't tracking. All that matters is that you placed 30 bids and won all of them. The probability of a given string of independent, random outcomes is equal to the product of the probabilities of those outcomes. In particular, the probability of winning a bet n out of n times is equal to the probability of winning it once taken to the nth power. Dinosaur is right: if the odds of getting red/black were 50/50, the odds of your winning streak occurring were less than one in a billion. I went ahead and calculated what the red/black odds would need to be for there to be a 50% chance of your winning streak occurring, and it turns out that the wheel color outcomes would need to be >97% predictable. For there to be a 1/1000 chance of your streak occurring, the color outcomes would need to be 80% predictable. Long story short, if you really won 30 out of 30 bets you placed, we are not talking about small biases.
good, you are starting to understand. Note that I did not say that there is no bias in the wheel, I said quite the contrary, that bias can NEVER be eliminated, just minimized. I already explained that, I have even given you two examples (the guys that didn't know what they were doing , having to struggle with a bias of \( -B \) vs. dr. Jarecki, who was "helped" by correctly understanding the bias). Once again: you are misquoting me, I said that the casinos minimized the wheel bias but that there are many other sources of bias. I also said that one can NEVER eliminate systematic errors in an experiment, your experimentalist colleagues know that as well. Err, I gave you a list, didn't I? "be it the depth of the shoe, the roundness of the ball, the way the croupier throws the ball, etc. " ...which is precisely what I said to Lakon way back when, the algorithm wasn't bad, it was just most likely backwards. Practically the reverse of Jarecki's algorithm , that won all the hands in a row, this one lost quite a few. I do not know how they did on more complex bets, they may have won <shrug> .
Wow. So way back on page 2 when you told me "there was no wheel bias", what you actually meant was "the bias associated with how the wheel spun was tiny compared to the biases associated with the shape of the ball, the surface of the wheel, the way the ball hit the wheel..." Meaning that when I said "the wheel itself would have to have an appreciable bias of some kind", your actual objection wasn't that there was no bias, but rather than the bias could have come from something about the wheel/ball system other than how the wheel was spinning. I really wish you had made that clear earlier. While looking up the exact betting rules for roulette, I found this Wikipedia section which I recommend to everyone. It doesn't include the Dr. Jarecki story, but it does describe a number of other, similar stories, including some in the surprisingly recent past. I enjoyed looking up further information on each case, and seeing how every group used a different technique to exploit a different kind of bias. With these possiblities in mind, let's consider Tach's case. What sort of bias were the programmers trying to exploit? Maybe it was a simple table bias, like the ones Gonzalo Garcia-Pelayo exploited to make his millions, such that the wheel favored certain pockets over others. (To be clear: I'm not saying that the spin of the wheel in particular favored certain numbers, just that something about the table did.) But if this were true, the programmers could have collected statistics about the wheel using one of those cards they hand out, then crunched the numbers offsite and returned. Bringing a computer into the casino would be needlessly clunky, and increase the odds that they'd be kicked out after succeeding for a while. Moreover, the (trivially) correct strategy to exploit such a bias is to find the most likely outcomes and bet on them every time, whereas Tach's programmers were changing their bets from run to run. We can conclude that the bias, if there was one, was not a simple bias of this kind. This leaves the kind of bias exploited by Claude Shannon and the Eudaemons (how's that for a band name?). In those cases, the gamblers waited for the ball to hit the wheel, then used the laws of physics to project its stopping point before placing their bets. The algorithms for such projections are still based on long-term observation of the wheel, but each projection has to be calculated separately and quickly, so the computer has to be there and the bets are constantly changing. Until today, I didn't know that people could place bets after the ball hit the wheel, so this changes things. Tach: were the programmers placing their bets before or after the ball hit the wheel? If the latter, I will now concede that it's possible, in principle, that Tach's strategy was actually a winning one. However, I still find it exceedingly unlikely, for the following reasons: 1. Tach was betting on the color and parity, which are the two values least correlated with real spatial quantities. Shannon/Thorp and the Eudaemons both succeeded by predicting roughly where the ball would land, then betting on numbers spread across the favored octant of the wheel. In other words, they calculated that the ball would whirl around by \(\theta\) radians before stopping, then bet that it would fall somewhere in the range \(\theta\pm \Delta\theta\) to account for experimental noise. By contrast, parity alternates every one or two slots and color alternates every slot, so betting on a color/parity instead of a specific number gives no added robustness to physically sensible noise. If betting on an octant is predicting \(\theta\pm \Delta\theta\), then betting on a color is predicting \(\theta\pm 4\pi n/37\) where n is some integer. If someone can come up with a noise mechanism that would make such a bet useful, I'd love to hear it, but as it stands I can't think of one. 2. The Eudaemons were the most recent users of this method mentioned on Wikipedia. Importantly, they applied it to the modern "low profile" tables, which are designed to make such predictions harder by lowering the angle of the wheel surface and the depth of the pockets. (Wikipedia says the low profile tables are ubiquitous now, so I assume that's what Tach was betting on.) They Eudaemons bet on octants and enjoyed a return rate of 144%, which translates to 72% prediction power. In order for a 30-win streak to be remotely feasible, Tach's method would have to predict the results with well over 80% accuracy. This means that not only would betting on colors/parities have to work, it would have to work much better than a recent best-effort to bet on octants. 3. Designing an algorithm that works involves two steps. First, you have to understand the physics of the system well enough to write the general form of the algorithm, with some free parameters. If you screw up this step, your predictions will be un- or weakly-correlated with the system's actual behavior, meaning they will perform as well as chance. Second, you have to make observations and pin down the free parameters, effectively "fitting" your algorithm to the system. In other words, you take some past results and adjust the knobs in your algorithm until it would have predicted those results. If you got the first step right, your newly tweaked algorithm would have predicted the past results for physically meaningful reasons, so it should predict future results with similar accuracy. Note that no matter how bad your model is, you will never perform worse than random sampling; step one determines how strongly your results are correlated with actual outcomes, and step two ensures that those correlations are positive. The only way to calibrate into a negative correlation is if the future results you're trying to predict obey different rules from the past ones you used to fit, which isn't true in this case because we're treating roulette spins as independent, identical events. Thinking through these reasons, I did come up with two scenarios under which I could believe Tach's story. In both cases, I assume the programmers came up with an amazing algorithm that could predict the exact winning number with high accuracy. Given such an algorithm, they might have taken the data from one day to calibrate it, then came back and tried to use it after the wheel had been tightened, the croupier had switched out, or some other change of variables had miraculously inverted the correct fitting variables. Alternatively, since Tach wasn't tracking anything except color/parity bets, they might have been winning big on specific number bets while placing low-stakes losing bets on the opposite color/parity to reduce suspicion and/or hedge their losses. Either would be an awesome story, so I hope one or the other is true. Apologies for such a long post - the possibility of betting after the ball drops just opened a big can of worms. Please Register or Log in to view the hidden image!
No one knows, so you are just fishing at this point. You can't conclude anything. You are simply fishing at this point. They did a lot of crunching after each turn, so they were making their bets very late, giving me very little time to place my bet. Yes, this is quite possible, like I said repeatedly, I wasn't watching their other bets. Te main reason was that they were placing the bets late, so I was scrambling to do the inverting. In addition, there was a lot of distraction produced by a diamond - bedecked lady, referred to as "madame la Comtesse" (the Countess), whose "strategy" was to place bets on all the squares.
Ok - here's what I'm going with. The algorithm was good, but they didn't want the house to know they were winning with it. So they predicted all the winning numbers, but split their bets between winning and losing so they came out about even. But while they were doing so, their inconspicuous companions would drift in and out of the game, responding to pre-arranged signals that told them whether the algorithm's bet would win or lose that spin. "Madame la Comtesse" was in fact on their bankroll, and had chosen an outfit/betting scheme to draw attention to herself and thus minimize the chance that someone would catch on to what was happening. Tach stumbled onto one of the simpler parts of their signaling scheme (whenever we make a color/parity bet, bet the other thing), and his 800 euro winnings were but a fraction of the hundreds of thousands the group took the casino for that day. And no one but Tach will ever know. Please Register or Log in to view the hidden image!
Nah, I think she was a fixture dating way back. Anyways, it was an interesting experience. If you go, you need remember that you need jacket and tie and that you need to pay entrance, this is not the main hall, it is one of the private rooms. Good luck!
Tach, I notice you haven't answered the above. I need to know this information, as I have a lot more to say, and a lot more scrutiny to subject your claims to. Extraordinary claims require extraordinary proof, wouldn't you agree ? This is a hard physics / math page - contributors posting anecdotes such as yours have been kicked off it for far less. You claims are highly implausable to say the least. Is there some reason why your claims should not be subjected to significant rigour and scrutiny ? Please answer the above. Also, you recently said to another contributor .. red/black and odd/even interspersed with a lot of other different bids that I did not participate in? Did you read the part where I told you that I stayed an undetermined number of rounds in-between? Your calculation is seriously off. Pleas advice; 1) What WAS the number of spins you stayed off bet within the hour where you DID make 30 bets ? You said undetermined, but surely, you must have some idea ? Just an approximation will do. 2) You assert above, that because you stayed off bet, that the contributors calculation of your 30 bets is seriously way off. Please explain why you think this. Remember, extraordinary claims .. extraordinary proof .. maths forum ..
There is nothing extraordinary in what I posted. The fact that you are unable to follow is your problem. Fednis48 understood ok. Riiight, so people trolling like you need to be kicked out for wasting server bandwidth and memory space.
Ockham's Razor; The principle that entities should not be multiplied needlessly; the simplest of two competing theories is to be preferred. You are a crackpot. And as I said much earlier on, you are basically a dishonest person. The simplest theory is that you had a thought bubble .. a delusion that some extraordinary event was happening. You were seeing ghosts. Your story rather reminds me of a John Travolta movie (I forget the name for the moment - was it 'Blessed' or something like that?) where something extraordinary happens to him, he develops psychic powers, then does some amazing things before dying of some brain problem. At the outset, like as in all similar movies, we are supposed to apply a liberal dose of credulity for a few moments and accept as a given the extraordinary, pseudoscientific thing that happened, and go with the narrative. It's all good from there. You repeated over and over that there was NO wheel bias. No roulette bias. No roulette bias. Would you like me to go through this and the other thread to show you haw many times you said this ? I had a suspicion that you WOULD defer to roulette bias to cover up your GIAGANTIC BLOOPER .. eventually .. that's why I went to great lengths to obviate this - see my post 2 on this thread amongst many others. You continued to confirm .. no roulette bias .. no roulette bias .. and then what ? ROULETTE BIAS ! Only now the roulette bias transfers from that of human perception, to that of perception by some wayward, reverse working algorithm .. useless .. priceless .. who knows ? This is NOT fit material for a hard maths forum where the burden of proof is on the claimant. This is synonymous to you saying initially .. I FOUND A GOLDEN EGG and then changing this to I FOUND A GOOSE THAT LAID A GOLDEN EGG .. BUT I CAN'T FIND IT NOW - GO FIND IT YOURSELVES. OTHERWISE, IF YOU CAN'T FIND IT, I'M RIGHT ! You repeatedly asserted that the algorithm was worthless (their algorithm wasn't worth anything) yet in post 91, you happily accept the possibility that it was (Yes, this is quite possible). In post 77, Dinasour worked out the odds of your claim at 2.44 billion to one. Not only are they accurate as far as they go, they are GROSSLY understated because they didn't go far enough - if looked at from the realistic point of view that you would have had to observe the magic algorithm for a number of spins immediately prior. How many times do you say you did THAT then ? 5 ? 10 ? 15 ? You won't say, will you ? Let us guess .. hmm .. I dunno, say 10. So, 40 spins in total, of meaningful information, which enabled you to assess, and then profit. Do some REAL maths for a change you GREAT CRANK, and calculate the odds of THAT! Does in go into the QUADRILLION .. maybe ? So, using Okhams Razor, what is the preferred theory ? Tach's fantastic magical algorithm thought bubble or odds of quadrillions to one ? Hmmmmm ? Funnily enough, there exist EXACT comments on your part in previous posts and threads, where you admitted it was all due to; a) luck b) intuition We'll get to those soon. For the moment, my algorithm tells me that the only goose round here, is YOU, though the egg you laid is by no means golden.
Putting aside the cause of the event in Tachs narrative, I think it is legitimate, essential in fact, in a maths forum of this nature, to try to quantify the actual odds of what he says happened. This is probably what should have been done at the outset. It is important to consider the number of meaningful events (spins). Tach mentioned 30. Unfortunately, he didn't provide this information until recently. When he did, contributor Dinosaur, put the odds of that at 2.44 billion to one. That’s good, but there's more to it. One group of events not dealt with, is the events (spins) that Tach observed which must have convinced him he was onto something. I mean, one doesn't just go blowing their hard earned money on gambling games, unless one is fired up by some pretty convincing evidence. Tach has not yet advised what those number of events were. Hopefully he will soon. In the meantime, we can speculate. I've been thinking this through very carefully. Earlier I speculated 10. I think 10 is conservative - it was probably closer to 15, considering human psychology, the degree of time that it would take a reasonable person to get over his incredulity, his fear of losing his money, etc, and to then start betting. But anyway, in the absence of further input from Tach, lets leave it at 10. Next, Tach said there was an undetermined number of spins when he didn't bet. Here too we have minimal information and must draw some reasonable conclusions. How long is a roulette cycle ? Electronic roulette tables that proliferate in clubs, run for about 30 seconds per cycle But in a casino, it is probably double that - say about a minute per cycle. So in Tachs hour at the Monte Carlo table, we can surmise that 60 events (spins) occurred. Lets be conservative and say 50. So Tach bet on 30 of those, but didn't bet on 20. Here it becomes murky. What of these 20 ? Were they not bet on at the advice of the algorithm, i.e., Tach following it's operators 'off bet' ? If so, they can reasonably also be counted as valuable information - a ‘DON"T BET’ signal is as valuable information as a ‘BET‘ signal. On the other hand, Tach did say he was distracted by some ‘madam’ or something .. I don't know .. he hasn't been very specific. However, I DO KNOW, and we can safely surmise that IF, as a consequence of distraction, Tach was NOT AWARE of what the algorithm was doing - whether those bets were winners or losers, we too can safely disregard them as inconsequential - even non existent. EDIT; So there is NO REASON in one case to exclude (subtract) those 'off bet' spins, and good reason in the other case, to ADD them to the total. So, pending Tach further clarification (I'm sure it's coming) lets make some more reasonable guesses. - 30 spins and wins as averred by Tach - 10 winning events prior to him starting to bet, that convinced him to go ahead. - Of the murkier group of 20 in the above paragraph, say another 10 relevant events. A total of 50 relevant events. I entered 1 - 50 in a column of 50 rows on my spreadsheet, and progressively multiplied by 2.1 - the .1 being the 5% house edge for the zeros. At row 30, I got 2,209,833,471 - 2.2 billion. That’s surprisingly close to Dinosaurs 2.44 billion. At row 40, the result was 3,685,975,827,806. Eeeeek !!! 37 trillion .. I think .. At row 50, I got 6,148,163,976,432,010. EEEEEEK !!! I have no idea what this number is - I would guess it's definitely way up in the quadrillion. Can anyone please help with this ? In an earlier posting, Tach tried to mitigate Dinosaurs 2.44 billion, saying that in amongst his 30 winning bets, he was off bet by some number of spins because of some distraction. But as I said earlier, if he was NOT conscious of those events, then for the purposes of the exercise they were non existent - nothing can be said about them. Thus, the realistic figure, is, must be, at least 40 significant events (I say closer to 50, but leave it 40) at odds of THIRTY SEVEN TRILLION TO ONE. So, whatever the cause of the events of that day, if we are to believe Tachs figures and extrapolate from those, they were off roulette wheel trend by at least 37 trillion to one. This brings about ENORMOUS problems no matter which story we believe - to be discussed later. In support of his claims, Tach has used terms such as experimentally verifiable, etc, but only recently has he provided us with SOME information for this to occur, and there is a good deal more outstanding. I would like Tach to provide more information - to fill in the blanks as it were, so we can subject them to more accurate scrutiny concerning their probability, as indeed should have been done at the outset. I have tried to attach my spreadsheet table below. Hopefully it will be there - if not, I’ll see what I can do. See EDIT, mid post. 1 1.00 2 2.10 3 4.41 4 9.26 5 19.45 6 40.84 7 85.77 8 180.11 9 378.23 10 794.28 11 1,667.99 12 3,502.78 13 7,355.83 14 15,447.24 15 32,439.20 16 68,122.32 17 143,056.87 18 300,419.42 19 630,880.79 20 1,324,849.66 21 2,782,184.29 22 5,842,587.02 23 12,269,432.74 24 25,765,808.75 25 54,108,198.38 26 113,627,216.59 27 238,617,154.84 28 501,096,025.17 29 1,052,301,652.86 30 2,209,833,471.01 31 4,640,650,289.12 32 9,745,365,607.15 33 20,465,267,775.01 34 42,977,062,327.51 35 90,251,830,887.78 36 189,528,844,864.34 37 398,010,574,215.11 38 835,822,205,851.73 39 1,755,226,632,288.63 40 3,685,975,927,806.12 41 7,740,549,448,392.85 42 16,255,153,841,625.00 43 34,135,823,067,412.50 44 71,685,228,441,566.20 45 150,538,979,727,289.00 46 316,131,857,427,307.00 47 663,876,900,597,345.00 48 1,394,141,491,254,420.00 49 2,927,697,131,634,290.00 50 6,148,163,976,432,010.00
I am not going to waste my time, server and memory bandwidth addressing each one of your long list of fallacies. You still don't get it, I did not bet against the house, I bet against a flawed algorithm. Fednis48 got it, why can't you? A such, all the childish calculations that you are doing are a total waste of time, the probability of winning is \(P'=1-p\) where \(p\) was the (very small probability) of the algorithm predicting the correct outcome, meaning that \(P'\) is orders of magnitude than the silly numbers that you keep producing, fact confirmed by experiment. This has been explained to you and to Dinosaur repeatedly. Yet, you keep repeating the fallacy of calculating the probabilities as if this was an equiprobable set of events, it isn't. This was a sure bet, as sure as betting with the correct algorithm of dr. Jarecki. This too, has been explained to you repeatedly. Must be the fact that being retired your cognitive abilities have gone significantly down and you have nothing better to do than troll the internet.
Trying to troll your way out doen't do it. A sure bet, huh ? 30 consecutive wins in a row on a roulette wheel. We should have heard about it in the annals of roulette gambling history. Or, casinos should have gone broke. We only have Tach's thought bubbles. A bit like Victors. You are so out of touch with the real world.
Nah, I simply pointed out your (repeated) fallacies for you. Please Register or Log in to view the hidden image!
