# Splitting Tens in Blackjack: Keep Your Hard 20

> Stand on 20, even against dealer 5 or 6. See a scoped EV comparison, card-counting exceptions, mixed-rank rules and the strategy lookup.

- Source: https://toolsgambling.com/blog/blackjack-split-10s
- Author: Evgeniy Volkov (ToolsGambling)
- Published: 2026-02-13
- Updated: 2026-10-08
- Cite as: "Should You Split 10s in Blackjack?", ToolsGambling, https://toolsgambling.com/blog/blackjack-split-10s

No—under ordinary basic strategy, you should not split two tens: keep hard 20 and stand, including against a dealer 5 or 6. Splitting creates two hands and requires an additional bet for the second hand. Card-counting deviations are a separate case, not the ordinary recommendation. This distinction is supported by [Nairn’s calculations](https://www.cof.orst.edu/cof/wse/faculty/Nairn/papers/splitting.pdf) and the [published Wizard of Odds index numbers](https://wizardofodds.com/games/blackjack/card-counting/high-low/).

## What to do against each dealer upcard

The current [basic-strategy tool](https://toolsgambling.com/casino/blackjack-strategy) returns **Stand** for two tens against all ten dealer upcard values under its six-deck default setup. These are results from the project’s model, not a claim about every blackjack variation.

| Dealer upcard | Model action for two tens |
|---|---|
| 2 | Stand |
| 3 | Stand |
| 4 | Stand |
| 5 | Stand |
| 6 | Stand |
| 7 | Stand |
| 8 | Stand |
| 9 | Stand |
| 10-value card | Stand |
| Ace | Stand |

The default rules are six decks; dealer stands on soft 17; doubling after splitting and on any first two cards; late surrender; US dealer peek for blackjack; and a 3:2 blackjack payout. Soft 17 means a total of 17 with an ace counted as 11.

The implementation uses an infinite-shoe approximation, recalculates for rules and applies documented deck-count overrides. It does not enumerate every possible outcome from a finite set of remaining cards. Separately, the [Wizard of Odds single-deck strategy chart](https://wizardofodds.com/games/blackjack/strategy/1-deck/) says to treat two tens as hard 20 when there is no separate tens-splitting row. Hard 20 is a total of 20 without an ace counted as 11.

## Why stand even against dealer 5 or 6?

Keeping a made 20 is better under ordinary basic strategy than breaking it into two unfinished hands. A weak dealer upcard may make two potential winning hands seem attractive, but the relevant comparison is the expected value of the actions—not the appeal of a possible outcome.

Expected value (EV) is the model’s average result per original bet under stated conditions. It is not a win probability or a prediction for an individual hand.

[Nairn’s paper, page 15](https://www.cof.orst.edu/cof/wse/faculty/Nairn/papers/splitting.pdf), provides an exact comparison for **one deck, 10-10 against dealer 6, with the dealer standing on soft 17**. The comparison is between standing with a maximum of one hand and splitting to a maximum of two hands. Values are expressed in units of the original bet.

| Decision in Nairn’s model | EV per original bet |
|---|---:|
| Stand on 20 | 0.697403 |
| Split into two hands | 0.525105 |
| Difference in favor of standing | 0.172298 |

For a hypothetical **\$25 original bet**, that EV difference is **\$4.30745**, rounded to the nearest cent as **\$4.31**. This is an expected-value difference, not a guaranteed loss from a particular split. These figures apply only to the stated single-deck model; do not apply them to six-deck games or other rules.

A successful split does not disprove the comparison. The decision is evaluated before the next cards are dealt, not by the outcome of one completed hand.

## Which cards count as tens, and can you split them?

A ten, jack, queen and king are each a **10-value card**, but equal point values do not mean equal ranks. Whether you can split a king and queen, for example, depends on [the table rules](https://www.pokerstars.com/casino/how-to-play/blackjack/rules/).

Some rules allow splitting cards of equal value; others require matching ranks. Do not assume that mixed-rank tens are eligible to split. For the basic-strategy decision, the important point is that two 10-value cards already make 20.

## When can card counting change the decision?

Splitting tens can be a card-counting deviation when the remaining card composition changes the relative EV of the actions. The [Wizard of Odds Hi-Lo Illustrious 18 table](https://wizardofodds.com/games/blackjack/card-counting/high-low/) publishes these index examples:

| Hand and dealer upcard | Index number for splitting |
|---|---:|
| 10-10 against 5 | +5 |
| 10-10 against 6 | +4 |

The source lists a six-deck liberal Vegas game with the dealer standing on soft 17, late surrender, doubling after splitting and resplitting up to four hands. This is a **different model** from Nairn’s single-deck example above.

The Hi-Lo count is a card-counting system that assigns +1 to cards 2–6, 0 to cards 7–9, and −1 to 10-value cards and aces. The cumulative total is the **running count**. The **true count** is the running count normalized by the estimated number of decks remaining, as explained in the [Wizard of Odds Hi-Lo guide](https://wizardofodds.com/games/blackjack/card-counting/high-low/).

An index number sets the threshold for an **index play**: a change from basic strategy within a particular counting system and rule set. The published +5 and +4 examples are not universal thresholds and should not be transferred to another game or true-count method. If you are not tracking the remaining card composition, these exceptions do not change the ordinary answer: keep 20 and stand.

## How to check two tens in the strategy tool

Open the [basic-strategy tool](https://toolsgambling.com/casino/blackjack-strategy), choose two tens and a dealer 6, and the current interface returns **Stand**. That means taking no further card and keeping your 20.

1. Use the visible card-picker buttons to select **T♠** and **T♣** as your cards.
2. Select **6♥** as the dealer’s upcard.
3. Read **Stand**, the displayed recommended action.

![English strategy lookup showing T♠ and T♣ for the player, dealer 6♥, and the recommended action Stand.](https://toolsgambling.com/media/article/2026/10/6710c657a2197e10901d24e95ae64195.webp)

*Local development interface showing two tens against dealer 6, captured at 390 CSS px with DPR 2. This is not a production-site or external-casino screenshot.*

The tool provides a basic-strategy action under its current model and settings. It does not calculate exact split EV for a finite set of remaining cards, incorporate your true count or verify the rules of an outside game. For this ordinary two-tens decision, its answer is clear: keep hard 20 and stand.

**Expected Value: Stand on 20 vs Split 10s**

|  | Stand on 20 | Split 10s |
|---|---:|---:|
| Dealer 2 | 0.64 | 0.47 |
| Dealer 3 | 0.66 | 0.5 |
| Dealer 4 | 0.69 | 0.54 |
| Dealer 5 | 0.71 | 0.57 |
| Dealer 6 | 0.7 | 0.56 |
| Dealer 7 | 0.77 | 0.4 |
| Dealer 8 | 0.79 | 0.34 |
| Dealer 9 | 0.59 | 0.16 |
| Dealer 10 | 0.55 | 0.13 |
| Dealer A | 0.6 | 0.07 |

## FAQ

### Should you split tens against dealer 6?

No under ordinary basic strategy: keep hard 20 and stand. The tool’s current model also returns Stand for two tens against dealer 6.

### Should you split tens against dealer 5?

No under ordinary basic strategy. A dealer 5 alone does not make splitting a made 20 the better decision.

### Can you split a king and a queen?

Both are 10-value cards, but whether you can split different ranks depends on the table rules. Check those rules before playing.

### Why do card counters sometimes split tens?

The remaining card composition can change the expected value of each action. Counters may use a true count and a rule-specific index number for an advanced deviation, rather than ordinary basic strategy.

### Does the tool calculate split EV?

No. It looks up a basic-strategy action in its current model; it does not calculate exact split EV for a finite set of remaining cards.