In 1910 Theo D’Alost published Théorie scientifique nouvelle du jeu de Roulette, Trente et Quarante in Brussels, presenting it as the result of roughly thirty years of study. The subtitle promised two laws capable of producing “hits against the bank” without progressions. Beyond the author’s emphatic claims, the work is important because it does not offer a simple martingale: it constructs an entire grammar of sequences, rapporteurs and transformations of the permanence.
To understand D’Alost, two opposite simplifications must be avoided. It is not correct to reduce him to a generic theory of “equilibrium and deviation”; nor is it correct to treat his claimed advantage as proven. The value of the treatise lies in the structure of its reasoning and its technique for representing the game, which we reconstruct here directly from the original edition.
The starting thesis: natural play returns to equilibrium
D’Alost opens the technical section with a severe criticism of ordinary systems. Playing the last winner, the avant-dernière, the two-spin pattern or any other fixed figure produces, in his view, the same structural outcome: over a long development, favorable and unfavorable forms tend to offset one another, while zero remains an uncompensated cost. He therefore considers both flat betting applied to an unchanging attack and progressions illusory, because they increase exposure without changing the generating mechanism.
His conclusion, however, is not that “nothing can be done”. D’Alost reframes the problem: instead of following natural production, one must construct a fictitious rapporteur, meaning a second representation of the same permanence in which the figures no longer retain the equilibrium observed in the original flow. The operational claim therefore arises from transforming the way spins are classified and aggregated, not from changing the physical probabilities of the wheel.
The core of Chapter III
In D’Alost’s language, natural play is a “vicious circle” of winning and losing spins. To escape it, equilibrium would have to be broken through an artificial rapporteur that distinguishes figures which stop more often at their usual extension from those which continue beyond it. These are the two points the author calls, respectively, the equilibrium point and the deviation point.
Figures: not merely Red and Black runs
The treatise classifies the permanence through figures: elementary forms derived from the succession of the two opposing chances. “One-spin”, “two-spin”, “three-spin” figures and so on represent different durations and configurations of runs and alternations. D’Alost does not assign absolute value to a single figure: he studies it in relation to the other figures and according to the way it is translated onto the rapporteur.
This approach is more sophisticated than simply chasing overdue outcomes. The author looks for relationships among forms, repetitions and stopping points. The question is not only how often a run appears, but how a figure develops, whether it ends at the expected point or extends beyond it, and what current it produces when recoded together with the others.
Figures and duration
The permanence is broken down into forms of different lengths. Meaning arises from the relationship among stopping, continuation and the opposite figure.
Rapporteur
A derived grid or sequence that translates the natural results. D’Alost uses it to make visible dominances that would be offset in the original flow.
Spin selection
Not every signal is played. The analysis is supposed to identify the spins belonging to the dominant current and disregard those that offer no advantage under the model.
Breaking and narrowing the game
The chapter devoted to “breaking equilibrium” develops a series of manoeuvres through which D’Alost attempts to produce increasingly selective rapporteurs. The concept of narrowing is essential: by stopping, translating or grouping figures, the author seeks to reduce the number of forms actually considered and therefore the “point d’écart”, meaning the size of the deviation the device must withstand.
Narrowing is not the same as a betting progression. Stakes should remain flat; what changes is the criterion used to decide whether a spin belongs to the observed mechanism. In theory, fewer exposed spins also mean less impact from zero. In the treatise, however, the ability to select genuinely favorable spins is assumed from the structure of the rapporteurs rather than demonstrated through a modern statistical protocol.
The eight figures and the regular/irregular pair
In the final section, D’Alost translates the figures into eight configurations. Figures 1, 3, 5 and 8 are defined as symmetrical; figures 2, 4, 6 and 7 as asymmetrical. In practice he calls them regular and irregular, abbreviated R and I. This pair becomes a new binary alphabet superimposed on Red/Black.
When figures repeat in runs, they form what he calls a “straight current” on the R/I rapporteur; when repetitions are arranged as alternations, a different current is formed. This step is important because it reveals the true nature of the method: D’Alost is not looking for an immediate periodicity in the colors, but builds successive levels of encoding. Each level transforms one permanence into another sequence, to which the concepts of run, alternation, equilibrium and deviation are applied again.
A professional reading of rapporteurs
A rapporteur is a descriptive instrument: it can highlight some structures and conceal others. For it to become predictive, the translation rule, entry signal, abandonment point and stake must be fixed before the data are seen. If the rapporteur that worked best is chosen after the event, an apparent advantage can easily be produced.
D’Alost’s position on flat betting, capital and zero
D’Alost insists that capital gives no mathematical strength to an unfavorable game. Fast or slow progressions may prolong the illusion, but a sufficiently long figure destroys the progression; meanwhile zero strikes increasingly large stakes. His “ideal attack” should therefore use flat betting, select only a subset of spins and obtain a margin wide enough to absorb zero.
This distinction is theoretically correct in form: stake management does not create expected value; information is needed that changes the conditional probability of the spins being played. The controversial point is that D’Alost regards his rapporteurs as capable of supplying that information, while the treatise does not present sufficient independent verification to separate genuine prediction, flexible selection and retrospective fitting.
The letters from Charles Henry and Cosmovici
The edition contains two letters often cited as scientific endorsement. Their meaning must be read precisely. Charles Henry calls the theory ingenious and says he considers it probably effective on the basis of observations by Cosmovici and Delmas, but states that he has not yet studied it thoroughly and refuses to lend his name as an authority. It is therefore not a mathematical certification.
Cosmovici is more favorable: he considers the introduction of regular and irregular figures, and especially the reduction of the point d’écart, decisive. He nevertheless concludes that experience will confirm the excellence of the system better than any reasoning. Even the most enthusiastic support therefore remains conditional on empirical proof.
What remains valid and what requires verification
- Still valid: the criticism of progressions as a source of advantage and the idea that selection must come before stake management.
- Historically original: the construction of rapporteurs and derived alphabets to analyze figures, stopping points and continuations.
- Requires proof: that a specific translation produces conditional probabilities different from those of an ideal Roulette wheel.
- Methodological risk: choosing among many rapporteurs, figures and stopping points after observing the permanence.
- Modern verification: frozen rules, a separate development sample, out-of-sample testing, inclusion of zero and comparison with independent simulations.
The professional interest of D’Alost’s work therefore does not lie in declaring him “successful” or “refuted”, but in reconstructing his actual object precisely: a theory for transforming and selecting sequences, designed to search for unbalanced currents without relying on betting progressions. It is one of the most elaborate architectures in the historical literature on Roulette, but its effectiveness remains an experimental question.
Theo D’Alost, Charles Henry and Armando Casuale
Master Armando Casuale’s Non-Pascalian Conception draws especially on D’Alost and Charles Henry. From D’Alost it takes the central role of representation, rapporteurs and the equilibrium/deviation dialectic; from Henry, the focus on the short term and the initial state. Casuale’s theory is, however, a later personal construction: it does not coincide with the two historical treatises and is not presented as a scientifically recognized theory.
Primary source examined in full: Theo D’Alost, Théorie scientifique nouvelle du jeu de Roulette, Trente et Quarante, digitized edition from Gallica — Bibliothèque nationale de France.