Sorry for the misleading title. But for most system players, it's not the house edge causing you to lose. Its variance, and how you deal with it. Suppose we had a wheel without the zero, with fair payouts. Most of the systems on here would still lose when negative variance comes in.
Let's say you walk up to a wheel (without zeros) and notice on the marquee that reds outnumber blacks. So you bet on black to catch up. But black may never catch up even if you played for millions of spins. The Law of Large Numbers only says that they will balance out statistically. But that doesn't mean that they'll even out. Even after a million spins, there could still be a difference of hundreds of hits between even bets.
Variance is our worst enemy, not the house edge. So, how do you deal with this? Many times flat bet may win. Sometimes a progression will be required (even on a non zero wheel). You have to be able to adapt when that negative streak from hell shows up...and it will.
Forget about the house edge. First, find a strategy that works on a non zero wheel.
Makes sense, if whatever method you're using can't produce over time on a no-zero wheel then it's a waste of time to keep tinkering with it.
House edge don’t matter 😱
Don’t say things that are forbidden in the forum.
Steve and the goons will come and attack
😂😂😂😂😂
I was actually curious and wondering the overall distribution of wins and losses at roulette.
House Edge is 1/37, ok.
You'd expect the sum, maybe average wins-losses of all players ever since roulette exists to be around -1/37. Ok.
What would the distribution curve look like? Would it be bell-shaped like SDs or wavy like the repeater probabilities graph, or just a straight line? Say x is the %age loss/gain and y the %age of players who ended at that win/loss percentage.
Of all the losers, at what % of their bankroll did they lose ( my guess most will be at 0%!).
Of all the winners, how in %age did they increase their bankroll ?
I guess it can be calculated...
Bigroben, that would be good info to know. I have seen it posted where the US casinos posted their actual take from roulette. The actual was different from casinos, but it ranged from about 15-25%. We would need more info to know what the real house edge is. I guarantee it is more than the mathematical house edge. I'm pretty sure casino could still profit nicely off roulette with a non zero wheel, that paid fair odds.
I totally agree, they feed people on every media with the "fact" that you lose because of the house edge. I'm sure BV is raking in huge profits on their no zero wheel. I personally have wiped out a few bankrolls on that game.
I'm sure that even if we removed number 36 and kept the payout at 35 to 1, 99% of the players would still lose.
Testing your strategy on a non zero wheel can be a real eye opener for some. Good money management is very important. Progressions may be necessary. BUT, if your strategy progresses from 1 unit to 100 units, that is way to steep! That's a $10 bet going to $1000 bets. Do people really play this risky like the simulations they post here? There are better ways to manage negative variance than extreme progressions.
You don't have to be profitable every session, as long as its profitable in the long run. If you can beat the non zero wheel in the long run, without resorting to steep progressions, then I'd call that a good strategy
The only possible downside is that it might hide weaknesses for longer than necessary that otherwise would be detected.
Pattern Breaker comes to mind as a perfect example, so many were playing (and winning in the short term) initially using it on BVNZ that it masked its true loss rate for too long before being exposed.
Still seems like a good starting point for the reasons you've mentioned Scarface.
Good thread with useful thoughts
I Always wondered if players can defeat roulette without the green 0.
I tend to say that it’s possible, look at what the probability say : on a fair wheel without 0, in the long term the player will win as much as the casino ..
Hence I keep saying that Roulette is a break even game for this very reason.
Then try to avoid variance with LWs strategys, sometimes can t avoid bad runs no matter what,
but many times yes. Why stay and bet on lossing streak when can avoid ? Attack when W s not LLLLLs ....
So yeah, playing on a fair wheel without zero is advantageous and can lead to profit, why ?
If you are sure that at some point you are going to break even, then you can attack the wheel and capture a successive / final win then you are winner
Quote from: Kattila on Jul 27, 05:22 AM 2018
Then try to avoid variance with LWs strategys, sometimes can t avoid bad runs no matter what,
but many times yes. Why stay and bet on lossing streak when can avoid ? Attack when W s not LLLLLs ....
Doesn't make any difference - just like that "perfect loser" nonsense quoted from Manrique. I've tested all these concepts comprehensively... each win/lose is independent from previous win/lose.
I was curious, suppose you had 2 wheels, one standard, one without a zero and played the same system for 150 spins on each. How many sessions would you lose, win or break even on each wheel? So I wrote a little program to find out. Here are the results of 100 sessions of 150 spins betting on High :
Winning sessions with no zero : 52
Losing sessions with no zero : 47
Break-even sessions with no zero : 1
Winning sessions with zero : 34
Losing sessions with zero : 58
Break-even session with zero : 8
P(W) on standard wheel : 34.00%
P(W) on NZ wheel : 52.00%
Since both systems are the same with the only difference being the house edge, and the probability of winning a session, denoted by P(W) is lower for the standard wheel, it shows that the house edge does matter. It's not really surprising because this is what theory says should happen. It doesn't matter if an individual session is short (in that case you would expect variance to dominate over the house edge), but as you play more sessions the house edge starts to take over.
This doesn't of course prove that no system can win, only that the house edge makes a difference, and a big one. I think you might be surprised at the difference in the results.
You can try putting in different values for the session length (SL) and number of sessions (NS). Go to
link:://rextester.com/l/pascal_online_compiler
then delete the code in the editor and paste in the code below. Press F8 to run the code and the results will be shown at the bottom of the window. Experiment with different values of SL and NS.
const
NS = 100; // number of sessions
SL = 150; // session length
var
z,nz : array[1..NS] of integer; // session results
i,j,spin,
zw,zl,bez,
nzw,nzl,
benz : integer;
begin
zw := 0;
zl := 0;
bez := 0;
nzw := 0;
nzl := 0;
benz := 0;
randomize;
for i := 1 to NS do begin
for j := 1 to SL do begin
// get spin (no zero wheel)
spin := random(37);
if spin in [1..18] then
dec(nz[i])
else if spin in [19..36] then
inc(nz[i]);
// get spin (standard wheel)
spin := random(37);
if spin in [0..18] then
dec(z[i])
else
inc(z[i]);
end;
// no zero session results
if nz[i] > 0 then
inc(nzw)
else if nz[i] < 0 then
inc(nzl)
else inc(benz);
// standard session results
if z[i] > 0 then
inc(zw)
else if z[i] < 0 then
inc(zl)
else inc(bez);
end;
writeln('Winning sessions with no zero : ':35, nzw:4);
writeln('Losing sessions with no zero : ':35, nzl:4);
writeln('Break-even sessions with no zero : ':35, benz:4);
writeln('Winning sessions with zero : ':35, zw:4);
writeln('Losing sessions with zero : ':35, zl:4);
writeln('Break-even session with zero : ':35, bez:4);
writeln;
writeln('P(W) on standard wheel : ':25,zw * 100 / NS:3:2, '%');
writeln('P(W) on NZ wheel : ':25, nzw * 100 / NS:3:2, '%')
end.
Joecoder,
Great stuff !
You hit the nails on the head !
That’s what matters !
Good stuff Joe! Do you happen to have the data for units lost and won? Just curious because wins and losses per session could be big differences. Let's say we ran 1000 test for 100 spins betting red only. Maybe results show an equal number of win/loss sessions...but end result could still end in a loss or gain of a couple hundred units. I'm pretty sure with the number of test you ran, it will be pretty equal to house edge
It does matter since if there was no zero you couldnt loose to zero
Quote from: Scarface on Jul 27, 07:49 AM 2018Do you happen to have the data for units lost and won?
I've added to the code so that it now shows the final balance, I assume that's what you meant? Sample run :
Final Balance (Zero) : -308
Final Balance (No Zero) : 105
Winning sessions with no zero : 49
Losing sessions with no zero : 47
Break-even sessions with no zero : 4
Winning sessions with zero : 38
Losing sessions with zero : 57
Break-even session with zero : 5
P(W) on standard wheel : 38.00%
P(W) on NZ wheel : 49.00%
New code :
const
NS = 100; // number of sessions
SL = 150; // session length
var
z,nz : array[1..NS] of integer; // session results
i,j,spin,
zw,zl,bez,
nzw,nzl,
benz,sumz,
sumnz : integer;
begin
sumz := 0;
zw := 0;
zl := 0;
bez := 0;
sumnz := 0;
nzw := 0;
nzl := 0;
benz := 0;
randomize;
for i := 1 to NS do begin
for j := 1 to SL do begin
// get spin (no zero wheel)
spin := random(37);
if spin in [1..18] then
dec(nz[i])
else if spin in [19..36] then
inc(nz[i]);
// get spin (standard wheel)
spin := random(37);
if spin in [0..18] then
dec(z[i])
else
inc(z[i]);
end;
// no zero session results
if nz[i] > 0 then
inc(nzw)
else if nz[i] < 0 then
inc(nzl)
else inc(benz);
// standard session results
if z[i] > 0 then
inc(zw)
else if z[i] < 0 then
inc(zl)
else inc(bez);
inc(sumz, z[i]);
inc(sumnz, nz[i]);
end;
writeln('Final Balance (Zero) : ':30, sumz:7);
writeln('Final Balance (No Zero) : ':30, sumnz:7);
writeln;
writeln('Winning sessions with no zero : ':35, nzw:4);
writeln('Losing sessions with no zero : ':35, nzl:4);
writeln('Break-even sessions with no zero : ':35, benz:4);
writeln('Winning sessions with zero : ':35, zw:4);
writeln('Losing sessions with zero : ':35, zl:4);
writeln('Break-even session with zero : ':35, bez:4);
writeln;
writeln('P(W) on standard wheel : ':25,zw * 100 / NS:3:2, '%');
writeln('P(W) on NZ wheel : ':25, nzw * 100 / NS:3:2, '%')
end.
House Edge with La Partage has been simulated for 60.000 placed bets - even money with various staking plans - many stay and end with positive results.
Is all in Philip Koetesh book.
Cheers
Ok, back to basics. If you're strategy can't survive a wheel with fair payouts, then it can't survive with house edge. Testing should be done with realistic bankroll. If test shows an aggressive progression like 1 unit to 100 units, and you know you'll never play like this with real money, scrap it. Try something different.
Even with a 1 million spin sample, reds can still outnumber blacks by 100s, even 1000. Flat betting a static system indefinitely will not work. There is nothing in the law of large numbers that's says ver selections have to equal out.
So, what doe that leave us with? The enemy of system players is variance. If your strategy can't be controlled on a non zero wheel, how can you expect it to do well when you factor in the house edge.
Negative variance will come. There is no way to avoid it. However, nothing says we have to make the same static bet, or follow the same rules that we make for ourselves.
What is in our control? How can you beat a wheel without zeros?
House edge is of great importance.
Why do you think why the betfair removed a few years ago wheel without a zero.
Because defeating such a circle is not a problem.
I even wrote a topic in this forum, where I described exactly how to beat a wheel without a zero.
But get a real edge of more than 2% and eve minimal low win over house edge, playing flat, or with minimal positive progression in long run is bordering on a miracle.
Quote from: MoneyT101 on Jul 26, 09:44 PM 2018
House edge don’t matter 😱
Don’t say things that are forbidden in the forum.
Steve and the goons will come and attack
😂😂😂😂😂
Flat earthers making jokes at the round earthers. Clueless.
Variance is not being understood here. Basically its the deviation from what you expect. There is always variance.
With the exception of increasing the amount of numbers you bet, you cannot reduce variance.
So if you want a low variance system, bet 30 numbers or so.
But it wont help if your win rate is still below payouts. Earth is round.
Variance is not the enemy. The odds or payouts arent the enemy. The enemy is the player's own ignorance.
are you You neglecting the effects of the house edge ?
Another way to reduce variance is to make more bets. As we know, in the short term anything can happen (meaning high dispersions), but these will even out as the number of bets increases. But the more bets you make the more you are exposed to the house edge grinding you down, so there's a trade off; isn't there always in life? ;-)
If you were able to lower the variance without having to make more bets or increase the quantity of numbers played so that you could use a progression safely, then you may still lose at the standard rate but would easily be able to make up the losses with your money management. Or you can attempt to increase your hit rate. Doing that may not reduce the variance but it would shift the bell-curve to the right and so away from negative territory.
So those are the two possible paths to the holy grail, as I see it. You don't have to get an edge in order to win, but lowering the variance may be just as hard to do.
Variance has no part in the hg because it alone doesnt address the fundamental problem, which is odds vs payouts.
Reducing variance still has a use if you want more predictable, and less volatile results. But again it still doesnt change the fundamental problem.
Quote from: Steve on Jul 29, 08:27 AM 2018Variance has no part in the hg because it alone doesnt address the fundamental problem, which is odds vs payouts.
Steve, why can't you bypass the problem of odds vs payout by reducing the variance? As I said, if you can lower the variance enough to use a progression or some clever money management without getting anywhere near the house limits, you can still make a profit even though you haven't changed the odds. And which is more important, changing the odds or making a profit?
Changing the variance isn't locked to changing the odds; I can't see any theoretical reason why you can't change the variance without having to change the odds. After all, you can do it the other way round; you can change the odds without changing the variance.
Quote from: Joe on Jul 29, 08:34 AM 2018you can lower the variance enough to use a progression or some clever money management without getting anywhere near the house limits, you can still make a profit even though you haven't changed the odds.
Thats not changing odds. Thats trying to win short term. Theres a big difference.
If your goal is to win most timesyou play, with the understanding when you do lose, you could lose all previous winnings and more, that's easy. Many systems do that.
But if you want an approach where simply the more you play, the more you earn.. then you absolutely must change the odds.
Quote from: Joe on Jul 29, 08:34 AM 2018Changing the variance isn't locked to changing the odds; I can't see any theoretical reason why you can't change the variance without having to change the odds.
Of course you can change variance. But odds are different.
You can decrease variance by betting more numbers. Technically this changes the odds BUT it does not change the difference between odds and payouts. When i say you must change the odds, i specifically mean you must increase accuracy of predictions to be high enough to overcome the unfair payouts. Its just simpler to say "change odds".
What you are talking about is finding an approach with more stable results.
The best approach for this is probably bet all the even chances for one side, and use a +1 progression after losses, and -1 after wins. Bet something like odds, blacks, 1-18. To lose, you would be a bad and rare losing streak.
I think the variance problem is tied up with the hold/cold fallacy. People think that the Law of the Third is a reflection of variance - but that's actually what we expect to happen.
Joe, I don't see how you can control the variance? Perhaps you could post a couple of examples?
The more numbers you bet, the lower your variance. Its really simple.
Of course prediction accuracy affects variance too but thats only for AP.
Actually i think the lowest variance you can get, and still profit, is betting 35 numbers.
You will almost always profit short term.
No matter what you do in roulette, if you take a shortcut, you need to pay for it somewhere. In this case, the price is small wins with possibility of large losses.
Reducing variance is really over rated. What matters is the discrepancy between odds and payouts. Or more specifically probability vs payouts.
One reason why gut and repeaters fail is they don't change probability. Why would the wheel care what it spun previously? Why would the probabilities suddenly change for you?
Don't get stuck in the "events must happen" nonsense. Sure, eventually black will spin after a streak of reds. But its meaningless. The probabilities still haven't changed. It's all in your head.
Once you understand the real problem with roulette is unfair payouts, you understand you must increase accuracy of predictions. Then you know to test a system, you need only test to see if its bet selection does that. Don't test over just a few spins. Test as much as you can. The more spins the better. An amateur response would be "but we will never play that many spins"
If you understand better, you know past spins have no connection to future spins on a random wheel. So turbos talk is waffle. He hasnt changed probability. And he isnt predicting hit numbers because there is no connection. Its very easy to disprove his hype but he just spoke trash when cornered. He still has some inexperienced players fooled.
Quote from: Steve on Jul 29, 09:00 AM 2018What you are talking about is finding an approach with more stable results.
I suppose you could put it like that but what it amounts to is reducing the variance. I'm not talking about doing that by betting more numbers, but reducing it for any fixed bet. For each bet in roulette there is a standard variance, but take any bet, say a dozen; would it be possible to reduce the variance for a dozen but not at the same time get an edge? The variance being smaller means less dispersion, which means
both shorter losing runs and shorter winning runs, so you haven't gained an edge. The expectation is still negative but you could make steady profits with a progression because it will never bust (assuming the variance is reduced enough).
These plots will explain it better I hope. This first one shows the distribution of wins and losses for two systems, green and purple. They could be any systems, but they both bet the same quantity of numbers, be it even chances, streets, whatever. The green system has a higher edge than the purple system because the mean of it is centred on 1, but the purple system's mean is centred on 0. That means you will break even playing purple and make a profit playing green. However the variance is the same for both (they are the same width) so bankroll fluctuations will be equally volatile, meaning the same size peaks and drawdowns.
(link:s://s8.postimg.cc/z6uhgs751/mean.png)
In this plot the mean is the same for both, zero. But the variance of the green system is higher than the variance for the purple system (the curve is wider). The mean is the edge so it means that you will break even for both systems but the volatility of the purple system is less than the green system.
(link:s://s8.postimg.cc/frf4a31dh/variance.png)
Quote from: falkor2k15 on Jul 29, 09:05 AM 2018Joe, I don't see how you can control the variance? Perhaps you could post a couple of examples?
Personally I'm sceptical, and I don't have any examples to give you. Some claim it can be done. But that's nothing strange in forums. ;-)
There's a thread by a guy named Nickmsi on the betselection.cc forum. He reckons it can be done and gives some examples.
link:s://betselection.cc/roulette-forum/use-mathstatistics-to-beat-roulettebaccarat-part-2/
In gambling, variance equates to your luck.
Claiming to control variance is like claiming that you can control luck.
Quote from: Joe on Jul 29, 12:44 PM 2018
Personally I'm sceptical, and I don't have any examples to give you. Some claim it can be done. But that's nothing strange in forums. ;-)
There's a thread by a guy named Nickmsi on the betselection.cc forum. He reckons it can be done and gives some examples.
link:s://betselection.cc/roulette-forum/use-mathstatistics-to-beat-roulettebaccarat-part-2/
He got it wrong - the true results are break even:
BBB 0
BBR 0
BRB -1
RBB -1
BRR 1
RBR -1
RRB 0
RRR 2
Quote from: The General on Jul 29, 01:25 PM 2018Claiming to control variance is like claiming that you can control luck.
I can control my luck by betting more numbers or betting for longer. ;-)
Reducing variance is not changing luck. It is just making results more predictable. I dont mean predictable like increasing accuracy of predictions. I mean your bankroll will not fluctuate as wildly.
Quote from: Steve on Jul 29, 03:29 PM 2018
Reducing variance is not changing luck. It is just making results more predictable. I dont mean predictable like increasing accuracy of predictions. I mean your bankroll will not fluctuate as wildly.
That's a good thing
Generally yes, but it still doesnt change your gradually eroding bankroll.
I wrote a little program which generates some stats based on the edge and how many numbers you bet. I've started a website where all my future programs can be downloaded from. There isn't much there at the moment but I will be adding stuff regularly over the coming months.
link:://:.roulettecoder.com/freeware.html
hi joecoder... joe the coder :xd:
:thumbsup:
Awesome joe, love these little quick programs. :love:
Learn to beat a wheel without zeros, or you'll never have a chance. When variance turns against you, what options do you have? Stop loss. Change bet selection/ system. Play less numbers. Do something different, or go down with the ship.
Playing more numbers with a steep progression is not the answer.
The only time the house edge doesn't matter is when you win more than you lose. Thats done by more accurate bet selection.
It can't be done with progression because thats just different size bets on different spins. Sometimes you get lucky and win back losses, sometimes you lose even more.
Avoiding variance is impossible.
Stop loss is just avoiding further losses by stopping play.
If your system works in the first place, the more you play the more you win, and stopping play is stopping profit.
a system working is relative
a stop loss can make winning systems for many people
How is winning relative?
How does stopping play make a system work?
Just curious of what most people think. If you were playing on a wheel, without zeros, with no house edge at all could you win long term?
Here are the rules: there are no rules. You have the right to switch bet selection...use progressions, stop loss, etc. Basically, you can choose or change your bet selection at anytime. You can use any money management you want.
Can you win long term on a wheel with no house edge?
I assume normal betting limits apply. Your bankroll will mostly hover at break even.
Sure if you use progression, the swings will be more pronounced, but youll end up the same place. Progression is a like a loan that eventually gets repaid.
Quote from: Steve on Aug 03, 09:27 PM 2018How is winning relative?
How does stopping play make a system work?
It works if you know the conditions of your play as it goes along. Like for instance you can know if you are experiencing a massive losing streak. You can know if you are in a win streak. Well, you should at least know those two things. Stopping while digging a hole is just common sense.
House edge is just an average that your wins and losses will tend to conform to. It won't force you to only have losing streaks. You can flat bet only, and target your win streaks by using a two level flat betting method. I know that it is not pure flat betting. It's not a progression either.
Maybe the math says it will break even. But its hard to believe that good money management would not help win long term on a zero edge wheel.
Quote from: Steve on Jul 29, 09:52 AM 2018Reducing variance is really over rated. What matters is the discrepancy between odds and payouts. Or more specifically probability vs payouts.
I agree completely. Variance is just there. The inequality of the payout is just there too. I can kill the casino with the very same thing that kills almost all system players. There is a sequence of death that takes out even the best devised systems where blind rules walk the player right into a perfect sequence that lines up just right in order to kill it. It does not matter much what the average loss rate is or even the fact that the payouts are not fair. To beat the casino you need to win more bets than you lose while playing even chance bets of 18 numbers selected to win. It's so simple. There is a way to do that too. It takes skill. Most people will say that that is impossible. Well that's good news for me. I hope they keep thinking that for years to come. There is such a thing as a sequence of life that equates to a sequence of winning.
Quote from: Steve on Aug 05, 06:18 PM 2018
I assume normal betting limits apply. Your bankroll will mostly hover at break even.
Sure if you use progression, the swings will be more pronounced, but youll end up the same place. Progression is a like a loan that eventually gets repaid.
I disagree. I think a player that uses a positive, up as you win, progression would fair alot better than a flatbetter or a negative progression player.
On a non zero wheel, the chance a flatbetter will win or lose 200 units will be 50%. A positive progression player will also lose 200 units 50% of time, but will win 500 or more units the other 50%. For the negative progression player, it will be the opposite...losses will be greater than wins.
Assuming all 3 players play the exact same strategy, for the same duration. If we were to look at 100 samples and compare, flatbetting will hover around even. Positive progression player will see greater wins than losses. Negative progression player will see greater losses than wins.
House edge is Alpha and Omega. Otherwise, Everyone can make perfect system.