My own system!

[ahlidap]

Basically I start fresh, spin the wheel, and wait until there are just 3 corners that haven't yet had a number hit in them.  3 unhit corners essentially.

Hi Mako,

Is number 2 one corner, or two corners?
Problem with the corners is that they aren't balanced.

1 corner is 4 numbers, but is impossible to cover the 36 numbers with 9 corners (this is what I mean by being unbalanced).

I used to use the first nr in the middle column for upper corner, and second nr for bottom corner. Example:
Nr 2 will be corner [2,3,5,6]
Nr 5 will be corner [1,2,4,5]

Is [5,6,8,9] a valid corner, or you play the "standard" ones only?

Have you tried to pick the first two corners that repeated?

[Bigbroben]

Robbert, Mako,
did a little test of 10k runs to check how many spins it would take to hit the coldest 1,2,3,4... corners.

Focusing on 2 corners, longest it took was 46 spins.
Distribution is as follows:

Expressed in percentage y hitting in x spins.
Overall, 80% hit within 8 spins, 99% within 24.


Maximum spins to hit, depending on the corners that were hit before, is:



Voilà.

[Mako]

Thanks Big Ben, exactly, any time I look at a "cold" process I look to the big spin sets that we use as a base for misses.  For quads/corners, the miss numbers are what you'd guess they would be, which is any 4 numbers (per quad).

Just like a full dozen can miss over 25 times, which is insane when you think about it, cold number systems tend to fail spectacularly when they fail to arrive over ridiculous amounts of time.

I hate them, any system that bets anything "cold", I discard, because I've never had one work under testing conditions with live spins.  For this particular one I let it pass, because we're resetting the target "cold" corners every say 25 spins, so the chance of hitting two of them, and those two being the coldest of the cold are probably pretty rare?

I like the idea though from davey, with the hitrate being as high as it is, I can see the testing ultimately producing an optimal balance on the progression, like say two full levels, if miss, then take the loss and reset to 111/222/44, etc.

That way even if you did manage to be "Mako lucky" and find the two quads that from that point forward are going to miss for 46 spins, you'd get off of them after X amount of attempts and cut your losses.

I didn't get a chance to do more runs yesterday, we'll see what the numbers say as we accumulate more data.

[Mako]

Hi Mako,

Is number 2 one corner, or two corners?
Problem with the corners is that they aren't balanced.

1 corner is 4 numbers, but is impossible to cover the 36 numbers with 9 corners (this is what I mean by being unbalanced).

I used to use the first nr in the middle column for upper corner, and second nr for bottom corner. Example:
Nr 2 will be corner [2,3,5,6]
Nr 5 will be corner [1,2,4,5]

Is [5,6,8,9] a valid corner, or you play the "standard" ones only?

Have you tried to pick the first two corners that repeated?

Yes ahlidap, exactly, there's some subjectiveness in how you pick those corners, what I've been doing is the following:

I wait for 3 full quads to be unhit, 12 total numbers.  I never allow overlaps, it has to be 3 full quads with 12 unique numbers that haven't yet arrived.  If I accidentally wait too long and there are only 2 full quads, that's fine of course, the only reason I wait for 3 full quads is that it's easier to achieve/start betting.

So no overlapping quads, each has to be a 4 number group.  But as you correctly mentioned, what about the number 2, or any central number that could be a part of two different corners?

In that case, you'll see it's rare.  Because the corners get eaten up by hits so quickly, that by 10 to 15 spins you won't be worrying much about multiple open corners and having to stress about which ones are the "right' ones.  Your options will be limited, and you'll say "ok there are 3 quads left unhit, I'm going to pick these 2 to start the method on" and the bet selection process for those 2 can be whatever you want (lucky numbers, location on the wheel-sectors or neighbors-etc, whatever) so long as you obey the rule that there are just 3x full quads/corners left unhit totalling 12 unique numbers.

So far that's worked, but the sample sizes are small.  I've done 500 spins on the trigger itself, plus the 306 spins on the trigger+Robbert's parachute, so while that trigger looks good more data may show it needs to be dialed in more, we'll see. 

As far as going the other direction, which I'm always a big fan of (hot corners being the trigger instead of cold corners), I would say definitely try it. 

It can't hurt, you'd have to create a definition for the trigger in terms of what a hot corner is, and is not, maybe any quad with 2 hits within 10 spins or something similar would be the definition of hot, but I would guess over tens of thousands of spins they would wind up roughly similar (betting unhit quads versus 'hot' quads as your trigger).

It's a nice way to play no matter how you play it, the foundation gives you a lot of options.  None of them may escape the house edge of course, but they may allow you to dance between the raindrops for longer than other methods for sure.

[nottophammer]

I hate them, any system that bets anything "cold", I discard, because I've never had one work under testing conditions with live spins.  For this particular one I let it pass, because we're resetting the target "cold" corners every say 25 spins, so the chance of hitting two of them, and those two being the coldest of the cold are probably pretty rare?

Rare happens
You must have read where talk of the missing doz for 5 spins posted by a programmer for casino. That wins in another 11 spins, total 16.
Well bruv got a mate who played this missing doz after 5 spins, betting 4 quads, so if middle column hit big win. He doubled from .25p.
It lost, then another game started and lost, the doz lost over 30 spins twice, so rare? oh yeah it was played on Smartlive

[Mako]

Rare happens
You must have read where talk of the missing doz for 5 spins posted by a programmer for casino. That wins in another 11 spins, total 16.
Well bruv got a mate who played this missing doz after 5 spins, betting 4 quads, so if middle column hit big win. He doubled from .25p.
It lost, then another game started and lost, the doz lost over 30 spins twice, so rare? oh yeah it was played on Smartlive

I believe it, you play long enough you see everything possible...mostly negative haha...but yes definitely. 

Progressions in my mind always have to stop/reset at some point to try and avoid catastrophic BR loses, the problem is that usually that progression stop winds up 'stopping' over time at exactly the house edge. 

Only an unlimited progression can exceed expected returns...right up until it hits the table limit and kills you, or an outlier 1 in 1,000,000 event occurs and kills you...

[Mako]

Another test, same parameters as my first but with a different spin set.

Now in this group we can see three large drawdowns, they occurred at these spins in the progression: 9, 10, and the largest, 12.

But those aren't the total amount of spins that the 2 corners/quads missed, as I'm not placing chips on them until there are just 2 or 3 remaining unhit quads. 

With those blank non-bet spins factored in, as if you had just played Robbert's method without waiting for the unhit quads, the total spins were 24, 21, and 21.

Those numbers are more in line with what Ben showed you, which is that two sets of four numbers can go missing for 45+ spins.  That's why you have two choices if you're going to base a play method off of something like this, either the Parachute/grind that Robbert suggested in his first post, or a more aggressive base betting unit, but with a built in stoploss/reset point in the progression as davey mentioned (111/222/44-STOP/RESET).

Now with the other test and this test we only have data on basically 40 overall betting attempts, and in those 40 attempts, 35 of them hit within the first level mentioned above (again, 111/222/44), while 5 of the attempts went longer (up to the 12th spin for the longest one I mentioned above).

If we had used davey's idea, the gross units gained equal +229u, while the losses for the 5 attempts that went beyond the first level total out at -90u.

Which nets out at a positive gain of +139u for the two sessions combined (716 total spins, 40 bets attempted).

We'd need at least 500 betting attempts (approximately 9000 spins worth) of data to really 'see' where the level progression cut-off should be, and if said progression cut-off is exceeding the house edge in terms of net units at that point.

It's worth pursuing, will continue to test a bit each day, I would encourage others to do so also.

[Roulettebeater]

I believe it, you play long enough you see everything possible...mostly negative haha...but yes definitely. 

Progressions in my mind always have to stop/reset at some point to try and avoid catastrophic BR loses, the problem is that usually that progression stop winds up 'stopping' over time at exactly the house edge. 

Only an unlimited progression can exceed expected returns...right up until it hits the table limit and kills you, or an outlier 1 in 1,000,000 event occurs and kills you...

Well
That’s a very interesting fact !
In roulette there is nothing called event is due in x spins !

Of course many believe that, but it’s clearly a sort of gambler fallacy ..

The question is what can the player do in order to win in this devil game ? What are the left options ?


I guess not much, but hit and run is an option, play the shortest time, hit and run, play one spin every day, only one ! If lost, keep your anger off and come the next day and play again a new spin...
What do you think of this approach ?

[The General]

Well
That’s a very interesting fact !
In roulette there is nothing called event is due in x spins !

Of course many believe that, but it’s clearly a sort of gambler fallacy ..

The question is what can the player do in order to win in this devil game ? What are the left options ?


I guess not much, but hit and run is an option, play the shortest time, hit and run, play one spin every day, only one ! If lost, keep your anger off and come the next day and play again a new spin...
What do you think of this approach ?

Equilibrium approaches are a foolish waste of time due to the gambler's fallacy.  Your only options are...

1.Visual ballistics
2.Wheel bias
3.Exploit sub par rngs

[Gitano]

Robbert, Mako,
did a little test of 10k runs to check how many spins it would take to hit the coldest 1,2,3,4... corners.

Focusing on 2 corners, longest it took was 46 spins.
Distribution is as follows:

Expressed in percentage y hitting in x spins.
Overall, 80% hit within 8 spins, 99% within 24.


Maximum spins to hit, depending on the corners that were hit before, is:



Voilà.

99% within 24....Good!! BigBroben could you bring available the excel for check the averages ?
If I wait 24th spins and start to bet the missing quads or singles numbers Do you think it 's a good idea?
Good system Robbert! Thank you

[Mako]

Another test, same parameters as my last one. 

As time passes I'm applying more of the averages that we have access to thanks to notto's near thousand full cycles to base my "entry" point at.

So for instance, lets say I'm at spin 12, there are just 3 unhit corners/quads remaining, which indicates that in those 12 spins there have been 0 or say just 1 repeat in total. 

I would hold off entering in that circumstance, because I would rather have at least 15 spins on record anyway, and more importantly, I want to see a few repeats 'catch up' before I start betting anything that's unhit at that time.

And conversely, during this run there were multiple times when there were 5+ repeats within the first 12 spins (including a run of 8 repeats within the first 13 spins haha), and in that case you want to badly jump in and bet because you know unhits are about to go on a run to 'catch up' to were the normal averages are in terms of unhits versus repeats in a given cycle, but you can't yet because there are 5 or 6 fully unhit corners/quads at that time.

In that case I wait until the bare minimum of 3 unhit remaining quads is present to jump in, as soon as possible basically, but never with more than 3 (or 3.5 if one quad is overlapped with another and shares 2 of the numbers) unhit before I actively place bets.

The end result of applying notto's (and to some extent Gordon's) excellent work on cycles to our unhit trigger is that it's lowering the amount of actual betting spins are needed before a hit, and providing a better picture of what you should, and shouldn't do, mechanically.

In this run the largest drawdown at spin 276 won on the 7th bet of the first level of the progression (111/222/44, winning on that first 4).  All of the other bets won at spin 4 or less, with an average betting spin total of just 2.7 for this particular spin set.

Too early to draw any conclusions, just fun to speculate on where the 'right' set of rules to engage in betting might wind up after a thousand of these triggers are done.

[Gitano]

Hi Mako! I just adore this trend in the graph! Really appreciate
could we have the .dgt for Rextreme software to each test by itself some own spins and share with you all / among us, the results?
Thank You

Gitano

[RouletteGhost]

Seems like too much work 

[-Katalyst-]

Equilibrium approaches are a foolish waste of time due to the gambler's fallacy.  Your only options are...

1.Visual ballistics
2.Wheel bias
3.Exploit sub par rngs

@Caleb! - Your "only" options are... blah blah blah 

*Let's play the ego trip here 
'Only'
- I don't think so!

....from my experience - there is a greater chance of profiting from this game playing  single numbers instead of the ECs  - granted there is a greater variance within the individual numbers (**need decent BR) as opposed to the outside bets but none the less - this is where it can be beaten

* there are systems that can achieve this *and no they are not raw mechanical systems


Nice thread Robert 

[The General]

Nope.  What's missing is "your experience".  History, common sense and math say that you are wrong. There aren't any systems that prove your point, and the mpr results are still dismal.