# How Do You Know the Chart Is Finally Learned?

A readiness checklist for basic strategy: accuracy that holds across hundreds of mixed hands, no stubborn cells, named rule sets, and survival across breaks and distractions.

Source: https://21trainer.app/articles/when-is-the-chart-learned/
Updated: October 7, 2026
Publisher: 21 Trainer, makers of Blackjack Strategy: 21 Trainer, an educational blackjack trainer for iPhone and Android (App Store: https://apps.apple.com/app/blackjack-strategy-21-trainer/id6759527695, Google Play: https://play.google.com/store/apps/details?id=app.blackjacktrainer)

> **Educational content, not gambling advice.** 21 Trainer and this article teach blackjack strategy and card counting as skills, using virtual chips only. No strategy or counting system guarantees winnings, and nothing here encourages real-money play. If gambling is a problem for you or someone you know, call 1-800-522-4700 (National Council on Problem Gambling).

Somewhere in the first weeks of practice, many learners have a moment: a session runs clean, the accuracy number looks excellent, and the question arrives. Is the chart learned? The honest answer is that a single good session cannot tell you. One strong block of hands is one sample, and a favorable sample is exactly what a partly learned chart produces best, because the deal spends most of its time on the rows you have seen the most.

The useful move is to treat "learned" as a claim you can test rather than a feeling you can have. Below is a readiness checklist. Each item is a test with a reason behind it, and the reason stands on its own: not "your accuracy went up," but why that particular kind of accuracy is the one that predicts performance at a table. Pass all of them and the claim is defensible. Fail one and you know exactly what to practice next.

The checklist assumes you are drilling decisions rather than just playing hands, the method our [practice guide](https://21trainer.app/articles/how-to-practice-blackjack/) builds its ladder on, and its thresholds come from pieces this site has already published.

## Accuracy that holds, not accuracy that peaked

The first test is the shape of your accuracy over time. A chart that is learned produces a floor: ordinary sessions, tired sessions, distracted sessions, all landing in the same narrow range. A chart that is not learned yet produces peaks: brilliant sessions interleaved with ones that fall apart the moment the deal turns awkward. The number to trust is the floor.

The thresholds this site has published are specific. The [four week practice schedule](https://21trainer.app/articles/blackjack-practice-schedule/) sets the checkpoint for each hand category at accuracy steady above roughly 95 percent across at least a few hundred hands, not a single lucky block, and it sets the bar for mixed drilling at table pace in the high nineties before a running count is added. Our piece on [what an accuracy percentage really means](https://21trainer.app/articles/blackjack-accuracy-percentage/) goes further and calls 99 percent or better across several hundred mixed hands the working target, while warning that any figure built on fewer than about 200 decisions should be ignored outright, because a short sample may never ask the questions you would fail. Both pieces agree on the core idea: the sample has to be big, mixed, and repeated before the number means anything.

## Mixed order, and then table pace

Accuracy measured in blocks is a different quantity from accuracy measured in mixed order, and only the second one counts. Our article on [blocked versus mixed drills](https://21trainer.app/articles/mixed-hand-drills/) explains why: a category drilled alone gives away the category. If every hand in the session is a soft hand, you never train the first decision a real table demands, which is recognizing what you are holding before the chart cell applies. An ace, a 4 and a 2 is a soft 17, and it is easy to misread as a hard 17 when nothing in the session has forced the distinction. Mixed drilling trains classification and recall together, and the schedule saves it for the final week.

Pace is the second half of the same test. Our piece on [decision speed](https://21trainer.app/articles/blackjack-decision-speed/) frames it as a diagnostic rather than a competition: a decision you cannot make in a couple of seconds is being reconstructed, not recognized, and reconstruction is exactly the process that distraction knocks over. Slow accuracy and fast accuracy are different states of knowledge. The chart is learned when the second one holds.

## No stubborn cells, and the rare rows are covered

Scan your misses across a few weeks of practice. If the same two or three hands keep appearing, the chart is not learned, whatever the aggregate says, because those cells have not landed. This is the argument of our piece on [keeping a practice log](https://21trainer.app/articles/blackjack-practice-log/): a short list of named, repeating errors is more useful than a record of chip totals, because each line points at a decision that can be drilled until it stops being wrong. In 21 Trainer, accuracy heat maps by hand type and dealer upcard, the Weakness Drill, which finds and drills your weakest hands, session history and advanced statistics are Pro features built around exactly this loop.

The rare rows are the quieter failure. Per our [starting hand frequencies](https://21trainer.app/articles/starting-hand-frequencies/) piece, in a six deck game hard hands make up 85.18 percent of the deal, each soft hand from A,2 through A,9 arrives 1.19 percent of the time, and a specific pair lands about once every 176 hands. Random dealing therefore under-tests the soft and pair rows, and the pair cells are also the most rule sensitive on the chart. Overall accuracy can look finished while the pair block quietly is not, which is why our guide to [drilling pair splits](https://21trainer.app/articles/drilling-pair-splits/) recommends forced repetitions for pairs instead of waiting for the deal to supply them.

## The rule set has a name

Ask which chart you learned. If the answer is "the chart," the chart is not learned, because there is no single chart to have learned. Our piece on [the hands that change with the rules](https://21trainer.app/articles/chart-changes-by-rule/) lists the cells that move: a small fraction of the total, but real, and concentrated in doubling, splitting and surrender decisions that deck count, the soft 17 rule and doubling after split all nudge. Knowing one chart is knowing one rule set, and the learner who cannot name it has no way to notice when the table in front of them answers a slightly different question.

The fix is the one our article on [practicing for a specific rule set](https://21trainer.app/articles/practicing-for-a-specific-rule-set/) prescribes: choose the rules first, drill exactly that chart until it is automatic, and only then go looking at how other rule sets differ. "Learned" is a claim about a named configuration, not about the game in general.

## It survives a break and a distraction

Two final tests measure whether the knowledge is stored the way you think it is. The first is time. Our piece on [coming back after a long break](https://21trainer.app/articles/relearning-blackjack-after-a-break/) describes decay as orderly rather than random: hard totals, the bulk of every session, are carved deep, while soft hands and pairs were always the shallowest rows and go first. A chart that was truly learned comes back much faster than it was first built, with the misses clustered in those shallow rows.

The second is noise. Our article on [practicing with distractions](https://21trainer.app/articles/practicing-with-distractions/) describes the gap between a hand answered in silence and the same hand answered with chatter, chips and a dealer waiting, and its method is to add the table's distractions to practice deliberately, one at a time, only after the quiet version is solid. Knowledge that holds only in a quiet room is real, but access to it can fail at a table. Learned means it survives both tests.

## What learned does not mean

Passing every item on the checklist still does not mean what most people assume it means. Perfect basic strategy reduces the house edge; it does not create a player edge. Under liberal multi-deck rules the edge against a perfect player is a fraction of a percent, losing sessions remain routine, and short run variance dwarfs anything strategy choices change within a session. The chart is the entry requirement for card counting, not a counting system itself: knowing it cold is the precondition for adding a running count without both skills degrading, which is why the practice ladder puts counting after automatic basic strategy, never before it.

That framing is also the healthiest way to use the checklist. Every item is a pass or fail with a named fix, which makes the whole question feel like inventory rather than hope. The free tier of 21 Trainer tracks accuracy, streaks and session goals, which covers the first test, and the Pro tier adds the heat maps, Weakness Drill, session history and advanced statistics that cover the stubborn cell test.

## Frequently asked questions

### How accurate should you be before the chart counts as learned?

Not the peak of your best session, but the floor of your ordinary ones. The checkpoints this site publishes are steady accuracy above roughly 95 percent in a single category across at least a few hundred hands, and accuracy in the high nineties on mixed drilling at table pace before any counting is added. A working target of 99 percent or better across several hundred mixed hands appears in our piece on accuracy percentages.

### Can you learn basic strategy in one weekend?

You can become familiar with the chart in a weekend, and that is a different thing from having it learned. Recall that survives needs repetition spread over days, mixed practice rather than blocked practice, and enough hands that the rare rows get tested. A single strong weekend session is one good sample, and one good sample cannot show that accuracy holds, that no cells are stubborn, or that the knowledge survives a break.

### Why does accuracy look finished while pairs keep going wrong?

Because overall accuracy is an average, and the deal is lopsided. Hard hands make up the bulk of the deal while each specific pair arrives roughly once every 176 hands and each soft hand about 1.19 percent of the time, so hundreds of correct hard total decisions can hide rows you barely ever see. The fix is to check accuracy by hand type, not in aggregate, and to give pairs and soft hands focused repetitions that random dealing will not supply on its own.

### Does knowing the chart mean you have an edge over the casino?

No. Basic strategy is the set of plays that loses least under a given rule set. Played perfectly it reduces the house edge to roughly half a percent under liberal multi-deck rules, and the house keeps that edge. Perfect play does not create a player edge, and losing sessions stay routine. The chart is the entry requirement for counting, not a way to win by itself.

### Is one chart enough, or do you need one per rule set?

One chart is enough for one rule set, and only for that one. Most of the chart is fixed by arithmetic that no rule changes, but a handful of cells move with deck count, the soft 17 rule, doubling after split and surrender. A learner who cannot say which chart they drilled is at risk of playing a handful of cells wrong all session at a table with different rules.

## Keep reading

- [What Your Blackjack Accuracy Percentage Really Means](https://21trainer.app/articles/blackjack-accuracy-percentage/)
- [A Four-Week Blackjack Practice Schedule](https://21trainer.app/articles/blackjack-practice-schedule/)
- [Blocked vs Mixed Drills: Why the Order of Practice Hands Matters](https://21trainer.app/articles/mixed-hand-drills/)