Random Thoughts

[falkor2k15]

Fair enough. Can't fault your reasoning there. At least it is confirming that your models are working correctly.

Since cycles have more stable ratios subject to less deviation, as well as bias, created out of the defining element linking one cycle to another, I don't have any expectation nor can I predict what will happen under various scenarios - nor what will be the best outcome - without first running tests. If I did then I would be deluded. And I don't trust my limited skills at maths to accurately predict probabilities theoretically. The computer simulation is far more accurate than me. When playing straights we would win before a cycle length of 25 - if there were no table limits because a repeat has to happen as per the Pigeon Hole Principle. So if we break that down into mini-games of CL1-4 via Quads then from what basis are you, bayes and turbo able to make such condescending remarks such as "it's to be expected" or "you are just proving probability works" or "you will never escape the house edge"? This isn't so much about escaping the house edge, but knowing the best outcomes in different situations and escaping the table limits whilst progressing towards a repeat that has to happen. Probability can tell us a lot - but it's overrated in the sense that a lot of you guys think it has the power to tell us that all systems will fail unless they use VdW. Similarly, people think that Derren Brown can read people's minds simply from using NLP as the sole explanation to his magical powers. The fact is: his shows are staged with stooges.

[RMore]

Wasn't trying to be condescending. Sorry you see it that way. Carry on my good man! (Trying to be lighthearted here!).

[falkor2k15]

But let's move on from there. Let's look at the numbers themselves - the ones that result from the transformation. It seems to me that there are actually two streams of data that can be analysed here. The series of numbers inside each string, and the flow of cycle lengths as each string completes. This makes it a sort of matrix if you will, a 2-dimensional animal that is writhing and straining against the constraints imposed by the cage in which we have placed it - but cannot ever escape. We may not ever be able to tame this beast but we can constrain it.

Rog, isn't there 3 streams here instead of 2...?
*Individual quads
*Cycle lengths
*Defining quads

Can we use the history of one to predict the fate of another? I hope to test that out in a couple of days time.

[falkor2k15]

Bad result again... 

When defining quad is quad 4 then which quad will hit on spins 2,3,4 (spin 1 is the defining quad 4 result carried over from the previous cycle)
Defining Quad 4 to Spin 2: 2750 (25.) 2850 (26.) 2794 (25.) 2727 (25.)
Defining Quad 4 to Spin 3: 2088 (25.) 2180 (26.) 2034 (25.) 1997 (24.)
Defining Quad 4 to Spin 4: 1066 (26.) 1066 (26.) 969 (23.) 1049 (25.)

[RMore]

Yes I guess so. You've added a new one - the flow of defining quads. My guess would have to be that this one would be pretty close to 25% for each quad and the flow would most likely look like any flow from a 25% probability. But in terms of how it relates to the other 2 flows? Mm - unsure of that. But on the surface, I would have to guess that it won't result in anything startling because I can't see any relationship between this one and the other 2. But of course it is only by digging into it that we can find out for sure. My advice though would be not to get your hopes up too high.

[falkor2k15]

Thanks - but I don't know how you can see a relationship between quads and cycle lengths - but by the same token - not between defining quads and cycle lengths or defining quads and quads? Without testing such hypothesis where does your logic/intuition come from? How are you able to instinctively recognise 1 combination being part of a transformation process - but not another?

[RMore]

Ugh - my brain is getting foggy! I think I need some sleep. I can see that there is a relationship between the quads contained in a string and the cycle lengths. If the cycle length is 1 then the defining quad absolutely must be the same as the cycle before. If the length is 2 then there is a 75% chance that the defining quad is the same as the one before - and so on. But I can't see how the flow of the defining quads, by itself, is related to either the cycle lengths or the quads contained in a cycle. But perhaps I'm just not seeing it. It's got nothing to do with intuition. It's got a lot to do with what I am capable of visualizing right at this moment.

[falkor2k15]

If the cycle length is 1 then the defining quad absolutely must be the same as the cycle before.

Sure - I can't function when I am tired!

"If the cycle length is 1 then the defining quad absolutely must be the same as the cycle before."
But by the same token:
If the cycle length is 1 then the quad must be a repeat of the previous quad before.
AND
If the defining quad is 1 then quad 1 must have repeated in the last 4 spins.
AND
RMore and Falkor both share an African grandmother!

So everything has a relationship - but how you are visualizing something significant between one combination over another needs a different explanation I feel. I guess dependency must exist between all of them to some extent? Of course it's knowing how to capitalise on that. Variance and the house edge always seems to win out in the end so am unsure yet how Priyanka's parallel games could be significant over others. Before, whenever I've tested dependency, it was different over different data sets: sometimes appeared dependent - other times gave a different result to what was expected. I wonder if within cycles we can find empirical evidence of consistent dependency across all data sets? If so would it result in edge or variance/deviation? The problem is: I could make the statement if stream 1 has such and such criteria then stream 2 has a 70% chance of such and such outcome. We can already say that 3 quads has a 75% chance of hitting. So it's difficult to understand. I guess the payout odds need to be better than the ratios - or the negative results to be outweighed by the positive. But that surpasses the limits of my intelligence (or knowledge?). Where can I get a crash course on why we can predict certain probabilities above 50% albeit without gaining any edge? Whose got some reading material for me??

Latest test:

Maximum Cycle Lengths in a row over 43763 cycles
Max Cycle Lengths in a row: 8 10 8 0

Cycle Length 4 result never came thru... must be a bug...

[falkor2k15]

Maximum Cycle Lengths in a row over 43763 cycles
Max Cycle Lengths in a row: 8 10 8 4

4 in a row!? That's hardly a dead-run situation is it? 

[falkor2k15]

Maximum Cycle Lengths in a row over 43763 cycles
Max Cycle Lengths in a row: 8 10 8 4

4 in a row!? That's hardly a dead-run situation is it? 

Where's bayes, general, iggiv or turbo to say "4 in a row - I could have told you that! But all systems lose in the long run; don't you know?" 

[falkor2k15]

OK it seems we can't use 1 previous cycle and 2 previous spins to predict whether we should play for CL2 or CL3 - CL2 seems to always be dominant. Priyanka must therefore decide to play for CL3 under some special trigger based around long term stats - perhaps to do with these parallel streams and associated dependency that has been alluded to.

Also, even though somebody told me to "look inside the cycle", it seems we cannot predict anything on a spin-by-spin basis - only by looking at the cycle event as a whole.

[falkor2k15]

I think I've figured out the best way to play this...

The VdW way is to pause after a loss; the non-VdW way is to carry on playing until the cycle is won...

We could begin our attack with the first bet on the defining quad - hoping to catch all those single length cycles - but as Rog says there's more chance of catching the other 3 quads instead (1 chip each street; zero left alone).

Once we get a win on the other 3 quads (first bet) this will then springboard us for an attack on the most likely Cycle closure: CL2. Here we can parlay our bets with 3 chips per quad and cover the zero too - 4 chips on the defining quad.

Long term stats should hopefully enable us to nail down with greater accuracy the more likely specific quads we should be targeting inside the cycle - and also when we should switch tactic to CL3 instead of CL2! Here we should bet slightly less than CL2: 2 chips per quad; 3 chips on the defining quad; 1 chip on zero.

In the rare events of losing to CL4 we should increase the span of the biased game by playing for 2 repeats or more! This part needs to be tested and figured out in terms of optimum play and correct weighting of bets.

[RMore]

One thing that I find really appealing about the quad cycles is that they are short. A loss is not bank-breaking. If there is an edge then progression should not be necessary. But even if there isn't an edge then perhaps a limited progression will do the business - on the basis that recovery should be fast since the cycles are short and the stats tell us that most wins occur on 2 and 3. Also that a good percentage do complete on the first spin and the outlay for that bet is just one quad (3 chips).

Just a little ambiguity with what you wrote though.

We could begin our attack with the first bet on the defining quad

Yep - got that.

Once we get a win on the other 3 quads (first bet)

Um - the first bet is on just one quad isn't it?

[falkor2k15]

Rog, I said I was considering the first bet to select just one quad (the defining), but having given much thought to it we should play more like Priyanka and based on your recommendation too: skip the first attempted closure on CL1 by instead betting the other 3 quads. Our attack on the cycle then begins proper on CL2 after getting a thrust from that first bet on the other 3 quads... soon we will be in orbit around Jupiter. I think that extending to 2-4 repeats will still keep the cycles short enough to not be caught by the law of large numbers?

[RMore]

Ah right. OK. Even though I have had a sleep I still seem to be mentally slow at the moment. Re-reading what you wrote I now see what you were saying.