Random Thoughts

[falkor2k15]

Developing a strategy from all these stats that are coming through is far from simple it seems... Which of these first 2 questions comes first - or do they both need to be addressed simultaneously?
1) Which Cycle Length should we play for next cycle?
2) Will the next cycle be defined by the previous quad or will it be different?

The next questions on from that will perhaps be:
3) How to best play for a such and such cycle length?
4) How to create further bias/edge during the above event.

Q1 depends on the best trigger for predicting the next cycle length - but also taking into account the best payout odds, which we know occur during CL2 and 3. How to evaluate constants and payout odds together? To understand if such triggers exist, more tests may need to be devised. Does ignoring CL1 contribute to this trigger or could there possibly be another reason why CL1 should be ignored?

Q2 could be tied up with Q1 because they are possibly dependent on each other? If you know whether the defining element will be the same or different then can you say more about the cycle length OR if you know more about the cycle length can you say more about the defining element in terms of "same"/"different"? This is the most confusing part for me.

Q3 involves playing for a set number of spins depending on which CL is being attempted, and there's different ways of playing those. Perhaps we should play only the last 2 quads instead of the last 3 or the last quad instead of the last 2? Do we include the defining quad or try to avoid it?

Q4 perhaps we could look at opportunities surrounding the last hit streets within the quads, or weight our bets differently across the quads, or look at extending the cycle to 2 repeats even. More stats needed to find out.

RMore, any feedback would be appreciated on the above!

[RMore]

It seems to me that all your hard work in the various ways of looking at the cycles is just confirming that internally the laws of probability are working as expected. There would seem to be no strange distortions occurring anywhere no matter which quad you choose to break down or follow. We must conclude therefore that if we want to develop a strategy based on this structure we must do so around the numbers that flow naturally from that overlay on the series of outcomes. And that's what this is, really. Simply an overlay on the flow of outcomes that transforms one series of numbers into another. Hence it is a transformation process.

But here we have imposed a carefully constructed structure such that it results in a series of discrete strings of numbers - that is, a string begins and ends rather than continues on endlessly.  Furthermore it is constructed in such a way that the string ends NECESSARILY rather than arbitrarily, so that while there is an endless supply of these transforms, each transform is complete unto itself.

Bottom line - I do think there is power in this type of transformation, but exactly how to capitalise on that is difficult and presently eludes me. Some would say this is because there is no possible advantage here no matter the "power" that I speak of, but I am not so sure. I think we do need to get a little creative and think outside the box a bit more. Simply running a series of statistical simulations is really doing nothing more than confirming that probability rules. The thing is, we already know this and so your effort in analysing the data, while admirable, is possibly wasted. Well - not entirely because I guess it has, at least, confirmed that probability is alive and well and functioning just fine here.

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.

Probably not much help but that's how I see it - at the moment anyway.

[falkor2k15]

(edit: haven't read RMore's latest reply yet!)

OK, I had a long think about this. I think 2 different strategies can be built up around both constants - initially beginning with one or the other: Cycle Lengths + Defining Element.

Defining Element strategy

In terms of Quads it seems the defining element is in our favour for each cycle to have more chance of finishing on it - even if you combined all the opposite quads against it:

When the defining quad is specifically quad 4 what chance does each individual quad have to close the next cycle (at Cycle Length X):
CL1: Defined by 4 to X: 0 (0.) 0 (0.) 0 (0.) 2727 (100%) = Defining element wins! 2727 vs. 0
CL2: Defined by 4 to X: 687 (17%) 724 (18%) 654 (16%) 1997 (49%) = draw! 2727+1997 vs. 687+724+654
CL3: Defined by 4 to X: 665 (22%) 726 (24%) 635 (21%) 1049 (34%) = Other quads win! 2727+1997+1049 vs. 687+724+654+665+726+635
CL4: Defined by 4 to X: 258 (25%) 255 (24%) 259 (25%) 273 (26%) = Other quads win! 2727+1997+1049+273 vs. 687+724+654+665+726+635+258+255+259 = 6046 vs. 4863 in favour of the defining quad!

So in this strategy we should aim to close off the cycle with the defining quad, gradually bringing in our knowledge of cycle lengths to assist us with decisions. I can envisage betting on the defining quad together with the most recent quad as a way of potentially gaining edge. Also, the fewer quads that we bet the better the payout.

Cycle Lengths strategy

I haven't yet found any support that a trigger exists for predicting the next cycle length, including missing out CL1, so why Priyanka does that is currently alien:

If previous Cycle Length is 1,2,3 or 4 then what percentage for the next Cycle Length to be 1,2,3 or 4
CL1 to 1: 2693 (25.) 2: 4010 (33.) 3: 3209 (28.) 4: 1007 (11.)
CL2 to 1: 4151 (25.) 2: 6055 (33.) 3: 4528 (27.) 4: 1562 (11.)
CL3 to 1: 3001 (24.) 2: 4676 (33.) 3: 3481 (27.) 4: 1216 (11.)
CL4 to 1: 1075 (26.) 2: 1555 (33.) 3: 1156 (27.) 4: 387 (11.)

However, the cycle lengths do have a natural bias to CL2 and CL3:

Cycle Lengths - overall general stats
CL1234 Totals 1: 10920 (25.%) 2: 16296 (37.%) 3: 12374 (28.%) 4: 4173 (10.%)

So we could build up a strategy of playing for Cycle Length 2 and 3 and introducing the other constant - defining element - to aid us with decisions a long the way. By stitching and parlaying bets we can get increase payouts. This is the strategy that Priyanka used in her videos, albeit flat-betting. However, she appears to be acting on triggers (if it finishes on one cycle length she ignores CL1 and goes for a repeat of the same) - but nothing has shown up in my stats to offer even an inkling as to the basis of those triggers!? 

[RMore]

Regarding why Pri leaves out CL1 - what if she does that out of necessity rather than choice? I mean, what if the bet strategy is such that there is no possible bet on the first spin? I dunno - just some random thoughts going on here. What could be a rule that results in that?

[falkor2k15]

Thanks Rog!  That's a step-up from my imagination of the ether and universal architecture!  Yours is indeed a beautiful analogy...  And I'll be following up on your advice pretty quickly... tomorrow I will see what I can come up with and hopefully take it from there? So then... until the next cycle of the earth... good-night my friend! 

[falkor2k15]

Regarding why Pri leaves out CL1 - what if she does that out of necessity rather than choice? I mean, what if the bet strategy is such that there is no possible bet on the first spin? I dunno - just some random thoughts going on here. What could be a rule that results in that?

OK I guess it's nothing more than a loss/virtual loss in the non-random world then - pausing play ahead of VdW to dictate the next action.

Next I am going to tackle the animal in the cage... anyone care to sponsor me?

[falkor2k15]

"But here we have imposed a carefully constructed structure such that it results in a series of discrete strings of numbers - that is, a string begins and ends rather than continues on endlessly.  Furthermore it is constructed in such a way that the string ends NECESSARILY rather than arbitrarily, so that while there is an endless supply of these transforms, each transform is complete unto itself."

Above Roger describes a "string" of numbers ending "necessarily" on a repeat, rather than stopping randomly; hence he is describing specifically the string of, say, dozens or quads, that make up a cycle; and these strings "transforms" into another string: cycle lengths.

"We must conclude therefore that if we want to develop a strategy based on this structure we must do so around the numbers that flow naturally from that overlay on the series of outcomes. And that's what this is, really. Simply an overlay on the flow of outcomes that transforms one series of numbers into another. Hence it is a transformation process."

"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."

Here again Roger describes how a string of dozens or quads, which make up cycles, are "transformed" into a 2nd stream of numbers/data that make up our string of cycle lengths.

"Cycle of lenght 1, cycle of length 2, cycle of lenght 1, CL1, CL1, CL2, CL3, CL2, CL4, CL3, CL1, CL2, CL1, CL3, CL3, CL3..... this is a sequence that is formed by events which are cycle lenghts."
link:://:.rouletteforum.cc/index.php?topic=17014.195

The "cage" is the framework of cycles imposed upon the dozen/quad strings, limiting their possible combinations into a set of events (4 different cycle lengths with 4 different defining elements) based on the pigeon hole principle, thereby putting a "constraint" on sequences of roulette spins that would otherwise continue endlessly or finish "arbitrarily". Roulette is a beast but we can constrain it using cycles - concentrating on their resulting streams of data - guiding us in controlling it's behaviour.

How to create a strategy from this? Well, just from looking at the newly created stream of cycle lengths it seems we could look for patterns amongst that stream alone, and hopefully expect more stable ratios of lengths that aren't subject to much variance - with the possibility of using a count to bring the lengths back up when one type has deviated? Also, there might be rare events contained in this stream that we could potentially capitalise on...

"...it is rare to have several unique 3 dozen cycles in a row.
I was tinkering around with the idea of betting for a repeat after seeing 3 unique fall.
As far as your idea here, how would we bet a sequence with repeats in the middle like this:

1233312222133321123

1233
33
3122
22
22
2133
33
3211
123

Just wondering how common it would be to have a string of unique 3 dozens with several repeats in the middle."
link:://:.rouletteforum.cc/index.php?topic=16937.30

Playing a system based around Cycle Lengths and/or the Defining Element - with or without VdW - our worst enemy is always going to be Cycle Length 4 (see above and below), resulting in "Dead-heats":

"While non-random is good, we often get into a dead-run. An example of a dead-run is below where you are trying to play for a dozen to repeat in 4 spins, you get sequences like 1231, 2311, 3121 etc. As Drazen and Turner rightly pointed out, there is still an opportunity to get these sequences over and over and over again that you can get into a deep hole. The key is how can overcome these dead-runs with a parallel bet or a parallel selection, which is the alternate game played on its own will give you a negative result, but played together will make this dead-heats into winning combination."
link:://:.rouletteforum.cc/index.php?topic=15938.90 (page 7)

However, ultimately, what Roger seems to be describing is that a relationship exists between cycle lengths and spins - both separate streams of data in their own right - that we can capitalise on:
"(Cycle Lengths) - That holds the key between differentiating from spins to events and the dependency it creates."
link:://:.rouletteforum.cc/index.php?topic=15938.690 (page 47)

Over the course of these topics we've analysed several different parallel games with supposed dependency existing between them. In (thinking) Outside The Box rrbb examined "transforming" ordinary roulette numbers into a 2nd stream of random numbers by carrying over the last result to the start of the sequence (a similar concept to the ending of a cycle and beginning of a new one by carrying over the defining quad); in the same context Herby examined "transforming" the rolls of a die into a 2nd stream of numbers (1-6) using the same framework:
link:://:.rouletteforum.cc/index.php?topic=17115.60

We've also examined "transforming" dozen/quad cycles into a 2nd stream of cycle lengths (countless times!) and even looked at VdW on ECs and Dozens simultaneously:
link:://:.rouletteforum.cc/index.php?topic=15938.0

What is this relationship between all these parallel games? And how can we possibly capitalise on it? I'm not even sure of the basics of how to play a 2 stream game of roulette over the single stream game that we are used to? It's not the same as playing at 2 different (independent) roulette tables because there is dependency between our two streams: "if one is peaking then the other is also peaking" (source: link:://:.rouletteforum.cc/index.php?topic=17014.225). Do we concentrate on each stream? Do we play one at a time or both together? Do we alternate at a set frequency? Or do we ignore the streams at large and concentrate on the "transformation" mechanism itself, i.e. the exact moment where a result in one stream is born out of it's counterpart? What are the basics we need to know in order to get us started on this?

Also, the problem as I see it is the nature of the "overlay" that occurs during this "transformation" process. On one side we have parallel games that compliment each other every spin - but on the other side are parallel games where the results occur at different points in each stream albeit with some overlap in the overlay:

OUTSIDE THE BOX - STRAIGHTS				VDW - ECS + DOZENS
Stream 1	Stream 2			Stream 1	Stream 2
27	27		28
4	5		26	Poss. 1,2,3 (High)	Poss. 1,2,3 (Dozen 3)
8	9		9
23	24		18	Poss. 3,4,5 (Low)
35	35		1
17	20		21		Poss. 3,5,7 (Dozen 1)
14	18		24	Poss. 1,2,3 (High)	Poss. 4,6,8 (Dozen 2); Poss.  6,7,8 (Dozen 2)
32	33		10	Poss. 3,4,5 (Low)	Poss. 8,9,10 (Dozen 1) Poss. 4,7,10 (Dozen 2)
9	15		11
13	19		12
OUTSIDE THE BOX - DICE				CYCLES + CYCLE LENGTHS
Stream 1	Stream 2			Stream 1	Stream 2
2	2		17	2
4	4		20	2	2
1	3		11	1
3	2		35	3
4	3		34	3	3
3	1		12	1
5	5		23	2
6	6		15	2	2

[falkor2k15]

(edit: haven't read RMore's latest reply yet!)

OK, I had a long think about this. I think 2 different strategies can be built up around both constants - initially beginning with one or the other: Cycle Lengths + Defining Element.

Defining Element strategy

In terms of Quads it seems the defining element is in our favour for each cycle to have more chance of finishing on it - even if you combined all the opposite quads against it:

When the defining quad is specifically quad 4 what chance does each individual quad have to close the next cycle (at Cycle Length X):
CL1: Defined by 4 to X: 0 (0.) 0 (0.) 0 (0.) 2727 (100%) = Defining element wins! 2727 vs. 0
CL2: Defined by 4 to X: 687 (17%) 724 (18%) 654 (16%) 1997 (49%) = draw! 2727+1997 vs. 687+724+654
CL3: Defined by 4 to X: 665 (22%) 726 (24%) 635 (21%) 1049 (34%) = Other quads win! 2727+1997+1049 vs. 687+724+654+665+726+635
CL4: Defined by 4 to X: 258 (25%) 255 (24%) 259 (25%) 273 (26%) = Other quads win! 2727+1997+1049+273 vs. 687+724+654+665+726+635+258+255+259 = 6046 vs. 4863 in favour of the defining quad!

So in this strategy we should aim to close off the cycle with the defining quad, gradually bringing in our knowledge of cycle lengths to assist us with decisions. I can envisage betting on the defining quad together with the most recent quad as a way of potentially gaining edge. Also, the fewer quads that we bet the better the payout.

Cycle Lengths strategy

I haven't yet found any support that a trigger exists for predicting the next cycle length, including missing out CL1, so why Priyanka does that is currently alien:

If previous Cycle Length is 1,2,3 or 4 then what percentage for the next Cycle Length to be 1,2,3 or 4
CL1 to 1: 2693 (25.) 2: 4010 (33.) 3: 3209 (28.) 4: 1007 (11.)
CL2 to 1: 4151 (25.) 2: 6055 (33.) 3: 4528 (27.) 4: 1562 (11.)
CL3 to 1: 3001 (24.) 2: 4676 (33.) 3: 3481 (27.) 4: 1216 (11.)
CL4 to 1: 1075 (26.) 2: 1555 (33.) 3: 1156 (27.) 4: 387 (11.)

However, the cycle lengths do have a natural bias to CL2 and CL3:

Cycle Lengths - overall general stats
CL1234 Totals 1: 10920 (25.%) 2: 16296 (37.%) 3: 12374 (28.%) 4: 4173 (10.%)

So we could build up a strategy of playing for Cycle Length 2 and 3 and introducing the other constant - defining element - to aid us with decisions a long the way. By stitching and parlaying bets we can get increase payouts. This is the strategy that Priyanka used in her videos, albeit flat-betting. However, she appears to be acting on triggers (if it finishes on one cycle length she ignores CL1 and goes for a repeat of the same) - but nothing has shown up in my stats to offer even an inkling as to the basis of those triggers!? 

I'm going to test some more stats tomorrow. The following stats were based on individual quads closing a cycle, but since the defining quad already has a head-start the other quads need to hit twice to overtake it and be the new defining quad.
CL1: Defined by 4 to X: 0 (0.) 0 (0.) 0 (0.) 2727 (100%)
CL2: Defined by 4 to X: 687 (17%) 724 (18%) 654 (16%) 1997 (49%)
CL3: Defined by 4 to X: 665 (22%) 726 (24%) 635 (21%) 1049 (34%)
CL4: Defined by 4 to X: 258 (25%) 255 (24%) 259 (25%) 273 (26%)

I actually need to test what chance there is of quad 4 and the other quads hitting at spin 2,3 and 4 - regardless of whether they close the cycle or not.

If the previous cycle length was the same then does that affect the next cycle length being same/different?
If the previous defining quad was the same then does that affect the next cycle having the same/different defining quad?

This statement is still a bit of an enigma:

"There is an important thing here around statistical advantage of same element defining the next spin. What if we remove cycles of length 1, do we see any difference in ratios. Can cycles of length 1 be exploited?  Can cycles greater than length 1 be exploited?"
link:://:.rouletteforum.cc/index.php?topic=15938.255 (page 18 )

Could Priyanka simply be referring to skipping the first bet and playing for cycle lengths 2 and 3 rather than missing out any previous cycles of length 1? We do get different ratios then:

Cycle Lengths - same or different to previous cycle length
CL1234 Totals (same): 12616 (29.%) (different): 31146 (71.%)

Cycle Lengths (CL1 ignored) - same or different to previous cycle length
CL234 Total (same): 13292 (30.%) (different): 30471 (70.%)

Cycle Lengths (CL1 ignored then 1st bet ignored) - same or different to previous cycle length
CL234 Total (same): 13292 (40.%) (different): 19551 (60.%)

In the first 3 Quads videos she does play the first bet - but it's not to complete the cycle. In the last 2 vids she missed out the first bet entirely.

[RMore]

Gosh that's a lot of writing there Falkor! This is really absorbing you isn't it? I really should look into Pri's vids a bit more. I seem to recall that she said that the first vid or perhaps 2 (?) she played arbitrarily but I think she also said at some point that certain ones were played accurately according to how she would actually play. Do you happen to remember which is accurate? And was there an accurate transcript of the outcomes and her play?

Sorry to ask because I know I could trawl through all the posts and probably find out for myself but honestly I am pressed for time and thought that of all the people lurking here you would most likely be the one to know.

[falkor2k15]

Rog, seeing as my day job of fixing computers only takes 5% of brain power, I need to keep my brain occupied with something more interesting! Could you please find me a better job? I'm a wasted talent.

I know that when Priyanka introduced the first quads video based on Iron Steel and Turner's ideas, she said it had a 9% edge - and that was the only video where an accurate transcript was posted before I came along and messed up that trend through being slapdash! 

Code Select Expand


Number Quad Cycle quad W/L Bet Why?
29 4
3 1
9 1 1 Bet 2 - 3 - 4 End of cycle: Bet all the other quads
26 3 W Bet 1 - 3 We won our first bet. Now we bet the last two quads
27 3 3 W Bet 1 - 2 - 4 End of cycle: Bet all the other quads
4 1 W Bet 1 - 3 We won our first bet. Now we bet the last two quads
27 3 3 W Bet 1 - 2 - 4 End of cycle: Bet all the other quads
32 4 W Bet 3 - 4 We won our first bet. Now we bet the last two quads
18 2 L No bet We lost. Wait for a virtual win.
1 1 No bet
7 1 1 Bet 2 - 3 - 4 End of cycle: Bet all the other quads
28 4 W No bet No bet. We wait for the virual win.
27 3 VL No bet Virtual loss.
24 3 3 Bet 1 - 2 - 4 End of cycle: Bet all the other quads
5 1 W No bet No bet. We wait for the virual win.
7 1 1 VW Bet 2 - 3 - 4 Virtual Win. End of cycle: Bet all the other quads
28 4 W Bet 1 - 4 We had our virtual win. Now we bet again the last two quads.
2 1 1 W Bet 2 - 3 - 4 End of cycle: Bet all the other quads
15 2 W Bet 1 - 2 We won our first bet. Now we bet the last two quads
31 4 L No bet We lost. Wait for a virtual win.
30 4 4 No bet ??? No ideal why we dont make a bet here…
14 2 VW No bet Virtual win.
29 4 4 VW Bet 1 - 2 - 3 End of cycle: Bet all the other quads
31 4 4 L No bet Here we lost our bet. Now we wait for a virtual win.
36 4 4 No bet
35 4 4 No bet
5 1 No bet
11 2 No bet
20 3 No bet
23 3 3 No bet
23 3 3 No bet
1 1 No bet
9 1 1 No bet No bet. We wait for the virual win.
27 3 Bet 1 - 3 Virtual win. Bet all the other quads. This bet is still active.
19 3 3 W Bet 1 - 2 - 4 End of cycle: Bet all the other quads
7 1 W Bet 1 - 3 We won our first bet. Now we bet the last two quads
15 2 L No bet Lost
10 2 2 Bet 1 - 3 - 4 End of cycle: Bet all the other quads
16 2 2 L No bet Lost
12 2 2 No bet
10 2 2 No bet
4 1 No bet
26 3 No bet
16 2 2 No bet
15 2 2 No bet
22 3 No bet
31 4 No bet
25 3 3 No bet
9 1 Bet 2 - 4 Virtual win. Bet all the other quads. This bet is still active.
11 2 W Bet 1 - 2 - 3 Here we see a new trend. Our previous cycle was lenght of 3. Now we bet it will also be 3. Bet the 3 previous quads.
23 3 3 W Bet 1 - 2 - 4 End of cycle: Bet all the other quads
25 3 3 L No bet
14 2 Bet 1 - 4 Here we switch bet. We now bet the two missing quads because we bet for a cycle of 3.
2 1 W Bet 1 - 2 - 3 Our previous cycle was lenght of 3. Now we bet it will also be 3. Bet the 3 previous quads.
5 1 1 W Bet 2 - 3 - 4 ??? Why bet? We did not have a virtual win here.
29 4 W Bet 2 - 3 We now bet the two missing quads because we bet for a cycle of 3.
20 3 W Bet 1 - 3 - 4 Our previous cycle was lenght of 3. Now we bet it will also be 3. Bet the 3 previous quads.
2 1 1 W Bet 2 - 3 - 4 End of cycle: Bet all the other quads
24 3 W Bet 2 - 3 We now bet the two missing quads because we bet for a cycle of 3.
16 2 W Bet 1 - 2 - 3 Our previous cycle was lenght of 3. Now we bet it will also be 3. Bet the 3 previous quads.
12 2 2 W END
link:://:.rouletteforum.cc/index.php?topic=15938.210 (page 15)

Apart from that, her most amazing video was the one on the Double Streets. I wish there were more videos of that, alas.

Next I have a very exciting test coming up ahead of the other ones I planned: which quad closes the cycle for CL 1,2,3, and 4?
*Most recent
*2nd most recent
*Oldest (defining quad)

[RMore]

OK. Thanks.

[falkor2k15]

Damn, that's a bad result:

Which Cycles are closed by which Quads in terms of most recent
CL1: always by defining quad otherwise cycle length has to be longer
CL2: last quad 8185 (50.) defining: 8111 (50.)
CL3: last quad 4105 (33.) 2nd to last: 4096 (33.) defining: 4173 (34.)
CL4: last quad 1069 (26.) 2nd to last: 1015 (24.) 3rd to last: 1015 (24.) defining: 1074 (26.)

[RMore]

It's to be expected. You are simply measuring a stochastic process in various different ways and so of course the results will be as predicted by probability. We are still inside the box.

[falkor2k15]

It's to be expected. You are simply measuring a stochastic process in various different ways and so of course the results will be as predicted by probability. We are still inside the box.

I know we are still inside the box, Rog, but which principles besides VdW are outside the box? I don't think Priyanka really shared enough info about parallel games, hence my post from yesterday to show the current state of affairs. The only reason I am performing these tests is because Cycles could be considered non-random and are based on the pigeon hole principle. And since Cycles are a new concept to us here they should first be put through the normal, conventional, tests first?

[RMore]

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