Most rows on a basic strategy chart are boring in the best way: a block of one action, then a block of another, with a clean boundary between them. The 9s row is not. It splits, then stops, then splits again, then stops for good, and the interruption sits at dealer 7 for reasons that look arbitrary until you see them.
Learn why that gap is there and you will never misplay the hand again, because the explanation is really a lesson about what splitting is for.
What the chart says
| Dealer upcard | Correct play with 9-9 |
|---|---|
| 2, 3, 4, 5, 6 | Split |
| 7 | Stand |
| 8, 9 | Split |
| 10, ace | Stand |
Two stands, seven splits, and the stands are not adjacent. That shape is what makes 9s one of the most commonly fumbled pairs, along with the plain fact that they arrive rarely enough that nobody accumulates natural repetitions, a problem our pair splitting guide covers across the whole pair block.
Start with what you are giving up
A pair of 9s is an 18. Not a great 18, but a made hand that beats a large share of dealer outcomes on its own. Splitting means abandoning it, paying a second bet, and starting two fresh hands from 9 each.
Nine is a genuinely good starting card, so those new hands are not bad. But the comparison is never "is 9 a good card?" It is "do two hands starting from 9 beat one hand of 18 against this specific dealer upcard?" Everything in the row falls out of that question, and the answer changes as the dealer's likely finish changes.
Against 2 through 6: split, because the dealer is in trouble
Against the weak upcards, the dealer is forced into the bust-prone script described in our hit-or-stand guide, breaking most often against 5 and 6. When the dealer busts, every hand you have on the table wins. Two hands collect twice.
That is the whole argument for splitting into weakness, and it applies to most of the pair block: you are not splitting because your new hands are strong, you are multiplying your exposure to a dealer failure that is coming often. Standing on 18 against a 6 wins one bet when the dealer breaks; splitting wins two.
21 TrainerPairs are rare at a table and easy to drill in an app. Instant feedback with the reason. Free on the App Store.
Get the app
Against 7: stand, because you are already winning
Here is the riddle, and the answer is one sentence: a dealer showing 7 makes 17 more often than any other total, and your 18 beats 17.
Think about what that means. You are holding a hand that is favored against the single likeliest outcome of the dealer's forced draws. Splitting would spend a second bet to walk away from a hand that is winning, in exchange for two hands that each have to build their way up from 9 and might well land on 17, 18 or worse. The dealer's 7 is strong enough to punish weak hands and not strong enough to beat your 18, which is a narrow window, and 18 sits exactly inside it.
This is also why the gap surprises people: the upcards on either side, 6 and 8, both call for splitting, so 7 reads as an inconsistency. It is the opposite. It is the one upcard where your made hand is precisely calibrated to the dealer's most likely finish.
Against 8 and 9: split, because 18 stops being enough
Move one card up and the calibration breaks. A dealer showing 8 most often finishes with 18, which ties your 18 rather than losing to it, and a dealer showing 9 most often finishes with 19, which beats it outright. Your made hand has gone from favorite to coin flip to underdog across two upcards.
Once standing stops winning, the comparison flips: two hands starting from 9 have a better outlook than one 18 that is pushing at best. So the chart splits again. Notice the logic is identical to the dealer 7 case, just with the arithmetic pointing the other way, which is why memorizing the row as a shape is worse than understanding it as a comparison.
Against 10 and ace: stand, because everything is losing
Against a 10 the dealer's likeliest finish is 20; against an ace, strength across the board. Your 18 is an underdog and so is anything you would build from 9s. When every option loses on average, the correct play is the one that loses least, and that is the one keeping a single bet on the table rather than funding a second hand into the same bad matchup.
This is the same principle that governs the game's ugliest hands: minimizing loss is winning, a point our article on expected value makes at length. Splitting here would double your money's exposure to a situation you are already losing.
How to actually learn it
The row compresses into one question you can ask at the table: does my 18 beat what this dealer is most likely to make? Against 2 through 6, the dealer often makes nothing at all, so multiply. Against 7, yes, so keep it. Against 8 and 9, no, so multiply. Against 10 and ace, no, and nothing else helps either, so keep the single bet.
Then drill it, because understanding a row and producing it under pressure are different skills, and 9-9 is too rare to learn from live play. A few targeted repetitions of the dealer 7 case in particular, per the approach in our drills guide, converts the chart's strangest row into one of its easiest.
Frequently asked questions
Do you split 9s in blackjack?
Usually. Standard multi-deck basic strategy splits a pair of 9s against dealer upcards 2 through 6 and against 8 and 9, and stands against 7, 10 and ace. That produces the chart's most distinctive shape: a run of splits interrupted by a single stand at dealer 7, then splits again, then stands.
Why do you stand on 9s against a dealer 7?
Because your 18 is already beating the hand the dealer most often makes. A dealer showing 7 finishes with 17 more than any other total, and 18 beats 17. Splitting would trade a hand that is winning against the likely outcome for two new hands starting from 9, which is a downgrade despite 9 being a good starting card.
Why split 9s against an 8 or 9 but not a 7?
Because the dealer's likely finish changes. Against a dealer 8 the common outcome is 18, which ties your 18 rather than losing to it, and against a 9 it is 19, which beats you. Once standing stops being a favorite, two hands built from 9 outperform one mediocre 18, so the chart splits.
Should you split 9s against a 10 or an ace?
No. Against a 10 or an ace the dealer's likely finishes are 20 and a strong hand respectively, so both your 18 and any hands you build from 9s are underdogs. When every option is losing, the cheapest one is the single bet: standing keeps one unit at risk instead of funding a second hand into the same bad situation.
Is splitting 9s affected by the dealer hitting soft 17?
The core pattern holds in both S17 and H17 games, and the 9s row is not among the cells the soft 17 rule commonly flips, though the exact composition-dependent margins shift slightly. As with every hand, use a chart computed for your actual rules rather than assuming a chart from a different game transfers.