Updated 2026-07-22 · 8 decks · S17 · DAS · methodology
9,9 (pair of 9s) vs dealer 5: the verdict
The math is in: split. Two hands of 9 beat one 18 against a dealer 5 — split them.
The numbers
| Action | Win | Push | Lose | EV per $1 |
|---|---|---|---|---|
| Split* | 54.4% | 6.9% | 38.6% | +0.363 |
| Stand | 53.9% | 12.2% | 33.9% | +0.200 |
| Hit | 17.9% | 2.6% | 79.5% | −0.615 |
| Double | 17.9% | 2.6% | 79.5% | −1.231 |
* win/push/lose shown per split hand; EV covers both hands per one initial bet.
Why split is correct
Played as a single total, this hand is weak against a 5. Split, each new hand starting from 9 wins 54.4% — enough that two of them (EV +0.363 across both) beat the best single-hand line at +0.200.
For context, a dealer 5 finishes on 17–20 about 47.5% of the time, on 21 10.8%, and busts 41.6%. Those dealer outcomes are the same no matter what you hold — the whole game is choosing the action whose distribution fits them best.
What the mistake costs
Choosing stand instead of split costs about $409 per 100 such hands at $25 a hand (−0.164 EV per dollar). The genuinely expensive error here is double: −1.594 per dollar, or roughly $3985 per 100 hands at the same stakes.
Does the rule set change it?
No — split holds across every common configuration: Vegas Strip, Atlantic City with late surrender, H17 games and European no-hole-card tables all agree here.
🎁 Blackjack Welcome Bonus — 500% + 70 FS18+ · New players · First deposit · T&C apply · Play responsibly Get this verdict at the table in one tap.Free Telegram mini-app: instant verdicts, drill trainer and rule-aware charts — no install, opens right in chat.Related spots
9,9 against other dealer cards:
Other key hands vs a dealer 5: