Deviations: When the Count Changes the Play

The idea

Basic strategy is computed for a neutral shoe. But you know something the chart doesn't: the count. When the remaining shoe is rich or poor enough in big cards, a handful of borderline decisions flip — the mathematically best play changes with the composition. A deviation, or index play, is exactly that: a basic-strategy departure triggered when the true count crosses a threshold.

The famous example

Insurance is the classic. Against a dealer ace, insurance is a bad bet in a neutral shoe — the book says never. But insurance pays when the dealer's hole card is a ten, and a high count means tens are exactly what's left. Rich enough, and 'never take insurance' flips to 'always take it.' That flip — the book's rule bending to the shoe's arithmetic — is the whole concept in one card.

Why we stop here

A working deviation set is a list of specific plays with specific true-count thresholds, learned as reflexes and drilled like strategy itself. That set — and the drills that grade it — is course material, behind the wall, where it can be taught completely instead of dropped as trivia. What belongs in a free guide is the concept, and the honest note that deviations are a refinement: the counting and the betting come first, and most of the money is there.

Why a play can change at all

Basic strategy is computed for a fresh, unremarkable shoe. When the composition of what remains has shifted far enough, the arithmetic that produced a cell can flip — the option that was second best becomes best. A deviation is nothing more than that: the same expected-value comparison, run against a deck that is no longer neutral.

So deviations are not a separate system or a set of tricks bolted onto the chart. They are the chart, recomputed for conditions the chart's own assumptions exclude. Which is why they only make sense to someone already playing basic strategy correctly and already tracking how the deck has moved.

Where the value actually is, and where it is not

Bet sizing does the heavy lifting; deviations refine the play. Betting more when the remaining cards favour you is the mechanism the advantage runs on, and changing individual decisions is a refinement layered on top of it. We measured the split rather than assuming it.

On identical shoes: the bet spread alone, playing no deviations at all, gains 1.392 percentage points over basic strategy. The full set of deviations alone, played at a flat bet, gains 0.049 — it moves a −0.503% ± 0.069 game to −0.453% ± 0.077, which is still a losing game by a margin no player would ever notice. Together they gain 1.560. **The spread carries 89.2% of the advantage; the deviations 3.2%.**

The two shares do not add up to the whole, and the part left over is the useful finding. Deviations are worth more when there is more money on the table, which is exactly when the count is high and the spread has already pushed the bet up. **They are not a second edge sitting beside the first — they are a multiplier on it.** That is the real reason they come after bet sizing rather than before: on their own they have almost nothing to multiply.

That ordering matters because it tells you what to learn and when. A player who has memorised a long list of departures but sizes bets poorly has put the effort in the wrong place. A player who sizes bets well and knows only the handful of departures that come up often is doing the substantial part correctly. More is better, and the errors multiply as fast as the refinements do.

The one worth knowing first

Insurance is the clearest case and the one taught first, because it is the departure with the largest effect and the simplest logic. It is a side bet on whether the dealer holds a ten underneath, and its value depends on exactly one thing: how rich the remaining cards are in tens. Basic strategy declines it always, which is correct for a neutral deck. When the deck is rich enough, that answer changes.

That is as far as a free page should go. The specific thresholds, the full set, and the conditions under which each one holds are course material — not because they are secret, but because a partial list applied at the wrong moment is worse than no list at all.

Learning them without breaking what already works

The failure mode is predictable. A player learns a handful of departures, starts watching for the conditions that trigger them, and their basic strategy degrades because attention is finite. The net result is worse than before, achieved through genuine extra effort.

The guard against it is sequencing. Deviations are added only once the chart is genuinely automatic — not merely known, but reflexive under distraction — and then one at a time, each drilled until it costs nothing to notice. A departure that has to be recalled is a departure that is stealing attention from the hand.

It is also worth being clear about what they are not. They are not a way to win hands that basic strategy loses; most of them are marginal decisions where the two options are close, which is precisely why a shift in the deck can tip them. The gain is real, cumulative, and small, and treating it as anything more is how it ends up costing more than it returns.

Why the free layer stops here

This page teaches that departures exist, why they exist, and where they sit in the order of things. It does not list them, and that is deliberate rather than commercial.

A partial list is worse than none. Applied at the wrong count, on the wrong rule set, or without the basic strategy underneath it being solid, a departure is simply an error with extra confidence attached. The full set only makes sense delivered with its conditions and drilled to the point of reflex, which is what a course is for.

Questions

How many deviations do I need to learn?

Fewer than you'd think — a compact set covers most of the value, which is why courses teach a curated list rather than hundred-row tables. Concept first, list later, reflex last.