I wish to bring to this discussion of 73 pages (!) Some precise notions to reduce the gaps on the dozens. The model of **palestis** (XXY + attack on 3 or 4 spins) is a very correct base. We are waiting for the model 2X + Y and we play 3 or 4 spins to hope Z. If Z comes, we will close the part always after the coming of a Z (example: // XYX (YYZ) //; // YXX (YZ) //; // XXY(YXX) XYYXYXYXZ //). It is important to wait until the end of the gap with the coming of a Z to close the part, that is what diminishes the deviances in the accountancy of the player. Of course, if // XYZ // the game is closed, we cannot play it. With this principle of play, we take spins that have a weaker deviance.

But we can improve this principle by using "*staggered games*". We take for example the sequence // XXY (YXX) XYYXYXYXXZ //. A first staggered will take 1 spin on 2 // **X**X**Y** (Y**X**X) **X**Y**Y**X**Y**X**Y**X**Z** //. We have here // **XYX (XYY) YX** //. It will be necessary to use shifts of 1 out of 2 and also of 1 out of 3, etc. We can thus follow several staggered games and play only the spins that indicate several times Z to play. The gaps will be further diminished.

My opinion on a negative progression is not favorable, because it is playing against the gap (too long). It must be emphasized that the negative progression assumes that we can beat roulette with the force of money. But the "black swan" will always remind us that **the only triumphant force against roulette is the discipline based on statistics**. Personally, I advise to follow this strategy in "*staggered games*" with unreal bet and of course to play according to the concept "*deviance-compensation*" (or another concept if players have a reliable knowledge). I know that players do not like a little complicated method. But it is necessary to make a choice, to enrich the casinos or to defend oneself correctly, with a technical level adapted to “*the reality of the laws of the hazard*”.