Playing Against a Dealer 3, Hand by Hand

The dealer's low upcards form two regimes with a border, and the border runs between the 3 and the 4. Against the 2 the chart is stingy, declining doubles and conditioning its splits, as we saw in the 2 column. Against the 4, 5 and 6 it opens up, the world of the 6 column where every stiff stands and the doubles come in bunches.

The 3 is the hinge. In the shoe game with the dealer standing on soft 17, it is the first upcard where the weak-card plays switch on, with hard 9, soft 17 and soft 18 all doubling and 6,6 becoming an unconditional split, and in every game it is the last of the dealer's weak upcards where hard 12 hits. Play the 3 like a 2 there and you misplay up to four cells; play it like a 4 and you stand on a 12 the chart still wants you to hit.

This piece walks the whole column total by total and pair by pair, then shows exactly which cells move when the rules change, which for this column is barely at all.

The whole column, one table

These are the strategy chart's cells for a dealer 3 upcard in the standard shoe game: four to eight decks, dealer stands on soft 17, dealer peeks for blackjack, blackjack pays 3:2, double allowed on any two cards, late surrender where offered.

Two notes before the reasoning. Nothing surrenders against a 3 in any chart, at any deck count, and a pair of 5s is simply a hard 10, doubling where the total doubles and never splitting. The rows conditioned on double after split, the 2s and 3s, are why splitting 2s and 3s and double after split reward reading over memorization, and the soft rows follow the general logic of soft hands: one card cannot bust any soft hand, so the weak ones, soft 17 or lower, never stand.

What switches on between the 2 and the 3

In the four to eight deck chart with the dealer standing on soft 17, four cells change between the 2 column and the 3 column: hard 9 and soft 17 go from hit to double, soft 18 goes from stand to double if allowed, otherwise stand, and 6,6 goes from a split that needs double after split to a plain split.

The pattern behind the four changes is a slightly weaker dealer: a 3 busts more often than a 2, and a double, which gives up the option to draw again for twice the bet on one card, is worth most against an upcard that busts often and finishes low. Hard 9's doubling window in shoe games runs from the 3 through the 6, and doubling hard 9 is entirely about where that window's edges sit. Soft 18 makes the same move and is easy to miss, since it looks like a finished hand; soft 18 deserves its own reading.

Scoping matters, though: hard 9 already doubles against a 2 in the one and two deck charts, 6,6 is already a plain split against a 2 there, soft 17 already doubles against a 2 in both single deck charts, and soft 18 already plays as double if allowed, otherwise stand, against a 2 in four to eight deck and two deck games where the dealer hits soft 17. The four-cell count above holds for that one chart.

21 TrainerFree on the App Store. Instant right or wrong feedback with a clear explanation on every hand. Get the app

Why the 3 plays this way

The numbers first. Per the six deck final total distribution tables computed by Wizard of Odds, with the dealer standing on soft 17 and already checked for blackjack, a dealer showing a 3 finishes on exactly 17 at 13.4 percent, 18 at 13.1, 19 at 12.5, 20 at 12.1 and 21 at 11.5, and busts 37.4 percent of the time. When the dealer hits soft 17, the bust rate edges up to 37.7 percent. Our dealer final total distribution piece lays out the full tables.

The stiff stands trace to that first number. Standing on a hard 13 through 16 is a bet that the dealer breaks, and against a 3 that bet pays 37.4 percent of the time, more often than the 2's 35.4 and less often than the 6's 42.3. Hitting those totals risks busting on five to eight of the thirteen ranks, and a player who busts loses no matter what the dealer does. From 13 up, the dealer breaks often enough that drawing costs more than it gains, so the column stands, the same arithmetic that holds the stiff stands together across the low cards.

Hard 12 is the exception, and the closest call in the column. Both hitting and standing lose on average against a 3, the published expected value tables put hitting narrowly ahead, and the chart hits, which is the whole story of 12 against a 2 and a 3. A 12 busts only on a ten value card, four of thirteen ranks, so the hit is cheap, and the dealer's 37.4 percent break rate is not quite enough to make standing the better loss. One upcard later it is: against the 4, with the dealer breaking 39.6 percent of the time, the same hand stands.

The doubles read the rest of the row. A dealer 3's made totals cluster low: 17 and 18 are the two most likely, at 13.4 and 13.1 percent, each higher total less likely than the last. A doubled 9 that catches a ten value card lands on 19, which beats a dealer 17, 18 or bust. A weak upcard, to a doubling hand, is one that breaks often and finishes low when it does not.

How the column moves with the rules

Barely, and that is the interesting part. The 3 column is identical in all six charts, four to eight decks, two decks and one deck, dealer standing on or hitting soft 17, with a single exception: 2,2. In both single deck charts, 2,2 against a 3 is a plain split, with no double after split condition. Everywhere else the pair needs DAS to split against a 3 and hits without it.

Contrast the 2 column, where the rules move hard 9, soft 17, soft 18 and 6,6 depending on the chart. By the time the upcard reaches the 3, the borderline cells have mostly tipped the same way in every game. The practical lesson is still the ordinary one: play from the chart computed for your game, because the exceptions live in exactly the cells that feel automatic.

Between the 3 and the 4, hard 12 goes from hit to stand in every chart, soft 15 and soft 16 from hit to double, and 2,2 and 3,3 are plain splits against the 4 in every chart, the single deck 2,2 already one against the 3. Those changes make the 3 the hinge of the low cards.

Practicing the column

Columns make good practice units: about twenty decisions against one upcard, small enough to hold in your head, varied enough to exercise hard, soft and pair logic together. Drill the 3 column on its own until every cell is automatic, then fold it back into mixed practice where the upcard arrives unannounced.

The errors worth hunting are the ones flagged here: standing on 12, hitting soft 17 instead of doubling, standing soft 18 on autopilot, and splitting 5,5 out of a vague sense that pairs split against weak upcards. In 21 Trainer you can set the deck count, the soft 17 rule and double after split in the rule settings, and the strategy chart redraws for the selection, so the cells you drill match the game you are preparing for. In Pro, the Weakness Drill finds and drills your weakest hands, and accuracy heat maps show whether errors cluster in these cells.

Drill the whole columnCharts that redraw for the rules you select, with instant feedback on every decision. No account, no ads, and your progress stays on your device. Download