Decision Library
Pair of 9s vs dealer 5 · six decks · dealer stands on soft 17 · double after split allowed
Correct play
SplitYou keep a positive expectation of +0.366 per unit here, and splitting beats standing by 0.165 per unit.
Deals a pair of 9s against a dealer 5 in the trainer, graded live.
Per unit bet, six decks, dealer stands on soft 17, computed by the engine.
| Option | EV / unit | vs. best |
|---|---|---|
| SplitBest | +0.366 | — |
| Stand | +0.200 | -0.165 |
| Surrender | -0.500 | -0.866 |
| Hit | -0.611 | -0.977 |
| Double | -1.223 | -1.588 |
Split a pair of 9s here rather than play the combined total. Each card starts a fresh hand against the 5, and two hands built from this rank do better than one stiff or mediocre total.
Splitting is worth +0.366 per unit; standing pat on the 18 would be +0.200. Playing two hands captures 0.165 more per unit.
Playing a pair of 9s as one hand feels safer, but it wastes the value of starting two hands from this rank. Hitting the combined total instead of splitting costs 0.977 per unit versus the correct splitting. Over a few hundred of these hands, that is real money handed back.
Same hand, different table conditions. The correct play holds unless noted.
| Table condition | Correct play |
|---|---|
| 6 decks, stands soft 17 (baseline) | Split |
| 6 decks, hits soft 17 | Split |
| Single deck, stands soft 17 | Split |
| Double deck, stands soft 17 | Split |
| No surrender offered | Split |
The play is stable: split is correct across deck counts, the soft-17 rule, and whether or not surrender is offered.
Strategy and expected values from a combinatorial engine validated against Wizard of Odds.