The Illustrious 18 & Fab 4: deviation charts

Deviations — index plays — are the specific spots where the true count overrides basic strategy. Basic strategy is optimal for a neutral shoe; when the count says the shoe is no longer neutral, a handful of borderline decisions flip. These charts are the complete list worth memorizing, and the sections below explain exactly why each play flips where it does.

Why a play "flips": the crossover

For any hand, each available action has an expected value that depends on the composition of the remaining shoe — which the true count summarizes. As the count rises, the EV of standing on 16 vs 10 rises (the dealer busts more) while the EV of hitting falls (you bust more). The two curves cross at some true count. That crossover is the index number:

Play A over B when EVA(TC) > EVB(TC)The index is the TC where the curves cross. 16 v 10: EV of standing overtakes hitting at TC 0 — so stand at any positive count.

Every index in the tables below was found the same way basic strategy itself was: exhaustive simulation at each count level, action by action. Nothing is judgment; each number is a measured crossover.

Notice the indices are two-sided. Positive indices activate an aggressive play as the count rises ("double 10 vs 10 at +4 or above"). Negative indices retreat from basic strategy as the count falls ("hit 13 vs 2 at −1 or below") — because a low-card-rich shoe makes your stiff hands safer to hit and the dealer's stiffs safer from busting.

The flagship: insurance at +3

Basic strategy says never take insurance, and for a neutral shoe that's provably right. Insurance is a side bet that the dealer's hole card is a ten, paying 2:1. Let p be the fraction of tens among the unseen cards:

EVinsurance = 2p − (1 − p) = 3p − 1Positive exactly when p > 1/3. A fresh shoe has p = 16/52 ≈ 30.8% — a 7.7% house edge on the side bet.

So insurance is a bad bet until tens make up more than a third of the unseen cards — and the count tells you precisely when that happens. Each point of Hi-Lo true count raises the ten density of the remaining shoe by roughly half a percent; the density crosses 33.3% at a true count of about +3. Above that, "never take insurance" is no longer advice, it's a mistake. This single index is worth more than any other deviation because a dealer ace shows up constantly and the flipped EV applies to your whole wager — which, at +3, is a big one.

Why only 18 (and 4)?

There are hundreds of computable indices. Don Schlesinger's analysis in Blackjack Attack showed the value is brutally concentrated: a small set of plays — frequent hands, big EV swings — delivers the large majority of all play-variation profit, and the long tail of exotic indices adds almost nothing while multiplying the memorization load and the error rate. HisIllustrious 18 (the core plays) plus theFab 4 (the four late-surrender indices) became the standard syllabus for exactly this reason: maximum EV per unit of memory.

The charts

Indices apply to multi-deck (4–8 deck) games and key off thetrue count. Each row overrides yourbasic strategy chart only when the count condition is met — at every other count, basic strategy stands. Where surrender appears, it assumes late surrender is offered; if it isn't, fall back to the basic strategy play.

Dealer hits soft 17 (H17)

Your handDealer showsTrue countPlayBasic strategy says
Any handA+3 or aboveTake insuranceNever insure
92+1 or aboveDoubleHit
97+3 or aboveDoubleHit
1010+4 or aboveDoubleHit
10A+4 or aboveDoubleHit
122+3 or aboveStandHit
123+2 or aboveStandHit
124-1 or belowHitStand
125-3 or belowHitStand
126-2 or belowHitStand
132-2 or belowHitStand
133-3 or belowHitStand
1410+3 or aboveSurrenderHit
159+2 or aboveSurrenderHit
1510-1 or belowHitSurrender, else hit
1510+4 or aboveStandSurrender, else hit
15A-2 or belowHitSurrender, else hit
169+5 or aboveStandSurrender, else hit
16100 or aboveStandSurrender, else hit
10,104+6 or aboveSplitHit
10,105+5 or aboveSplitHit
10,106+4 or aboveSplitHit

Dealer stands on soft 17 (S17)

Your handDealer showsTrue countPlayBasic strategy says
Any handA+3 or aboveTake insuranceNever insure
92+1 or aboveDoubleHit
97+3 or aboveDoubleHit
1010+4 or aboveDoubleHit
10A+4 or aboveDoubleHit
122+3 or aboveStandHit
123+2 or aboveStandHit
124-1 or belowHitStand
125-3 or belowHitStand
126-2 or belowHitStand
132-2 or belowHitStand
133-3 or belowHitStand
1410+3 or aboveSurrenderHit
159+2 or aboveSurrenderHit
1510-1 or belowHitSurrender, else hit
1510+4 or aboveStandSurrender, else hit
15A+2 or aboveSurrenderHit
169+5 or aboveStandSurrender, else hit
16100 or aboveStandSurrender, else hit
10,104+6 or aboveSplitHit
10,105+5 or aboveSplitHit
10,106+4 or aboveSplitHit

A warning about misapplied deviations

Every index play is a bet that your count is right. Deviate on a wrong count and you've made a play that's worse than basic strategy in the actual composition you're facing — you'd have been better off knowing nothing. This is why deviations are the last layer of the learning path, not the first: their value is conditional on flawless execution of everything underneath.

Frequently asked questions

Why is insurance the most valuable deviation?

Two reasons: frequency and size. A dealer ace appears in roughly 1 in 13 hands, far more often than any specific hand-vs-upcard combination, and the EV swing when insurance flips positive applies to your full bet. Simulations consistently attribute more value to the insurance index than to any other single play — often a third of all play-variation profit.

Do deviations change between H17 and S17 games?

Mostly no — the indices are nearly identical. The differences concentrate around a dealer ace, because an ace is more dangerous when the dealer hits soft 17. Use the chart matching your game; both are published above.

Should I learn deviations before I can keep a reliable count?

No. A deviation applied at the wrong count is a basic strategy error with extra steps — it subtracts EV. The order is: perfect basic strategy, then an error-free running count and true count conversion, then deviations. Each layer is worthless without the one below it.

How much are deviations worth compared to bet spreading?

Bet spreading is the majority of a counter’s edge; play deviations add roughly 10–20% on top in typical shoe games, more in single- and double-deck where the count swings harder. They also matter tactically: standing on 16 vs 10 at high counts protects exactly the big bets you just pushed out.