Deviations: The Hands Where the Count Overrules the Chart
Updated
Basic strategy is the best play when you know nothing about the shoe. Once you are counting you do know something, and a handful of decisions change. Most of the value is in one of them.
- Insurance at +3 is worth more than every other deviation combined.
- The rest are worth a fraction of a percent each.
- Index numbers are not exact constants — they shift with rules and deck count.
- Applying one at the wrong count is worse than not deviating at all.
What a deviation is
A basic-strategy chart assumes an average shoe. When the count tells you the shoe is ten-rich, some borderline decisions tip the other way — standing on 16 against a ten becomes right, insurance becomes a good bet, doubling 10 against a ten starts to pay.
The count at which a decision flips is its index number. The set most people learn is the Illustrious 18, chosen because those cells come up often enough to matter.
| Hand | Basic strategy | Deviate to | At true count |
|---|---|---|---|
| Insurance | decline | take it | +3 |
| 16 vs 10 | hit | stand | 0 |
| 15 vs 10 | hit | stand | +4 |
| 12 vs 3 | hit | stand | +2 |
| 12 vs 2 | hit | stand | +3 |
| 11 vs A | hit | double | +1 |
| 9 vs 2 | hit | double | +1 |
| 10 vs 10 | hit | double | +4 |
| 12 vs 4 | stand | hit | below 0 |
| 13 vs 2 | stand | hit | below −1 |
Insurance is the one that matters
Insurance pays 2:1 and needs the dealer to have a ten underneath one time in three. On a fresh shoe that happens 30.9% of the time, which is why basic strategy declines it.
At a true count of +3 there are enough tens left that the real figure crosses 33.3% and the bet turns profitable. It is worth more than every other deviation put together, because it is a whole extra bet rather than a small shift in one decision.
Index numbers are softer than they look
Published index numbers are printed as though they were constants. They are not: each one is a crossing point that depends on the deck count, the soft-17 rule, and how the shoe actually got to that count.
We tested this by computing our own crossing points with an idealized shoe. Eight of the fourteen hands we checked matched the published index exactly; the rest came out one or two points higher. That is not a contradiction — it is what happens when you change the assumption about how a shoe reaches a given count.
Indices only pay if they are automatic, which means drilling them rather than reading them. Our deviation trainer is free and runs in the browser.
The practical lesson: treat an index as a guideline, not a constant, and do not agonize over being one point off. The calculator on this site uses the standard set.
When deviations cost you money
Applying a deviation at the wrong count is worse than never deviating. If your count is unreliable — and early on it will be — the deviations amplify the error rather than the edge.
The order to learn things in is: basic strategy until it is automatic, then an accurate count, then insurance, then the rest. Skipping to the deviations because they feel advanced is the most common way people lose money while believing they are ahead.
Keep reading
By Martin Kosek · every number on this site is computed by us, not copied.
← Back to guides