Because I was using the same stake on zero in those examples. A straight-up number pays 35:1, so with a $2 bet the total return is $72, including the original $2 stake. The screenshots were only a few examples to illustrate what I was talking about, not the full set of observations.
I see what you mean, and that is actually very close to the distinction I am trying to make. I am not really interested in zero simply as the end of a dry spell, but in whether the relationships around it retain some structure when the reference point is changed. Using random non-zero hits as controls is useful precisely because some visually convincing patterns should fall apart once you apply the same criteria elsewhere. Where my interest goes a little further is that I am not looking at a single “shape” by itself. I am paying more attention to how several conditions behave together within the same surrounding sequence, and whether those relationships remain stable across different samples. So yes, I agree that 200–300 zero events, while separating cases where zeros occur too close together, would make the comparison much cleaner. If the difference disappears under that test, that tells us something. If it remains, then it becomes much more interesting.
That makes sense. I agree that the variables and the window should be fixed beforehand, and that the same 12/12 structure should be applied to zero and to control numbers such as 17 and 32. I would also test nearby window sizes, because if the effect only appears under one very specific range, I would be much more cautious about treating it as meaningful. The same applies to overlapping zero events; I would separate those cases so they do not distort the comparison. What I am really trying to isolate is whether the relationships I am observing around zero remain consistent when the same rules are applied elsewhere. If they also appear with similar frequency and structure around ordinary reference points, then zero is probably not special. But if the difference remains across a larger sample and under the same fixed criteria, then I think it becomes worth examining more closely. That is why I am more interested in the stability of the relationships across samples than in any single sequence or isolated occurrence.
Yes, I agree that the windows should be defined before looking at the result. Otherwise, it is very easy to end up selecting the section of the sequence that best fits the hypothesis after the fact. Using 12 spins on each side could be a good starting point, although I am also interested in checking whether the effect — if there is one — remains when the window size is changed, rather than depending entirely on one specific choice. I also think it is important to use control points that are not zero and apply exactly the same measurements to them. That is where you can really start to distinguish between a structure associated with the event being studied and normal fluctuation in the sequence. As for sample size, I completely agree. A short sequence by itself proves almost nothing. My interest comes precisely from accumulated observations, not from one or two isolated examples.
I understand your point, and in fact that is one of the first things that needs to be ruled out. My interest does not come simply from the fact that zero stands out visually, but from comparing whether certain behaviors keep the same structure when the reference point changes. I am not assuming that zero “causes” what happens before or after it. What interests me is checking whether certain configurations around it show a different distribution from what you get when using ordinary spins as reference points. I have already noticed some differences that seem consistent enough to justify looking deeper into them, but for now I still prefer to treat them as observations rather than as a rule.
For quite some time, I have been conducting my own research, collecting data and studying different types of behavior that can be observed in roulette. One of the subjects I have been paying particular attention to is what happens around the appearance of zero . My main interest is not necessarily trying to predict exactly when zero will appear. Instead, I focus on observing what kind of behavior was developing before zero appeared and what begins to develop afterward . Over time, I have compared different sequences of spins and found certain behaviors that I consider interesting. For example, I don't simply look at how many spins have passed without a zero. I also pay attention to how other characteristics of those spins are developing and whether certain sequences or configurations tend to appear more frequently around those moments. One thing I find particularly interesting is that zero does not necessarily have to be studied as an isolated event. It may be more useful to look at the overall state of the sequence at the moment zero appears , what had been happening several spins beforehand, and what happens during the spins that follow. I'm not presenting this as a mathematical rule or as an infallible method for predicting zero. It is simply one of several lines of research that I have been studying for quite some time. My intention is to determine whether certain behaviors repeat consistently enough to deserve further analysis. I would be interested to know whether any of you have made similar observations: Have you noticed any particular behavior that tends to occur before or after zero appears, beyond simply counting how many spins it has been absent? I would also be interested in hearing about your own research or patterns you may have discovered over the years, particularly observations that are not usually discussed in traditional roulette systems.