Advanced research build — not certified for live play. The 280-cell chart matches its documented reference matrix, but strategy EVs and simulator model accuracy are not fully certified. Do not treat outputs as betting advice or a promise of advantage.
♠
Configurable 1–8 decks · Hi-Lo · S17/H17

Blackjack Strategy Lab

Strategy reference, live count, and shoe-level simulation
Local app · saves count on this device

Open finite-shoe research workbench · Experimental; exact state-limited and Monte Carlo modes are separately labeled.

$500 bankroll$5 minimum3:2 blackjackDASLate surrender75% penetration2 split hands by defaultDouble hard 9–11 only
Your configured game

Six-deck American hole-card blackjack

Default settings are six-deck American hole-card blackjack. The simulator can vary decks, dealer soft-17 rule, split limits, bankroll and cut-card penetration.

Starting bankroll$500
Blackjack payout3:2
Dealer soft 17Stands
DoubleHard 9–11
Quick start

Use the tools

Check a hand, track exposed cards, or run a Monte Carlo comparison.

Live count snapshot

Running count
0
True count
0.00

Cards seen: 0 · Decks remaining: 6.00 · Floored TC: 0

Important guardrails

A positive count does not guarantee a winning session. Index values are reference deviations, not a rules-certified optimal strategy for every casino. The Kelly strategy requires externally validated edge and variance assumptions; the interface does not validate them.

Simulation results are estimates from the model and are not promises of future performance.

Hand decision helper

Enter your two cards and the dealer upcard. View the workbook-derived reference for research. Hi-Lo deviation recommendations are disabled pending independent validation.

HIT

Enter a hand to get a recommendation.

Double restrictions

Doubles are permitted only on hard 9, 10, or 11. The app converts an otherwise-ineligible double recommendation to a hit.

After a split, doubling is allowed on eligible hard 9–11 totals. No resplitting is allowed. Split aces receive one card each.

Hi-Lo card tracker

Enter each exposed card once. Do not enter the dealer's hidden hole card until it is revealed.

Running count
0
True count
0.00

Cards seen 0 · Estimated decks left 6.00 · Floored TC 0

Recent cards

No cards entered yet.
Decks remaining are estimated as 6 − exposed cards ÷ 52. The true count is a practical estimate, not an exact unseen-card calculation.

Five-strategy comparison

Each strategy uses independent reproducible seeded sessions. Configure decks, S17/H17, DAS, penetration, table maximum and bankroll. Set rounds per seed and the number of independent seeds. Confidence intervals use a Student-t interval across independent seed/session means; model error is not included.

Checking whether the selected simulator rules match the fixed six-deck S17/DAS strategy reference…
Kelly is disabled until all three assumptions are supplied: estimated edge at TC 0, edge change per +1 true count, and variance per unit wager. This app does not validate those assumptions; do not use guessed values. The sizing output is sensitivity analysis only, not a betting recommendation.
No simulation has been run in this session.

Strategy definitions

A. Basic strategy, flat $5.

B. Hi-Lo tracked, flat $5, no index deviations.

C. Disabled pending independent I18/Fab 4 EV and rule validation.

D. Disabled pending independent I18/Fab 4 EV and bet-ramp validation.

E. ¼-Kelly sensitivity model with a configurable cap; available only for explicitly supplied assumptions. It is not a betting recommendation and its assumptions are not validated by this app.

What the simulator reports

Rounds played, total initial wagers, total action, net result, Return on total action, ending bankroll, maximum drawdown, win/loss/push counts, and whether the bankroll was exhausted.

Bankroll-depletion frequency is the observed fraction of independent sessions that stopped below the table minimum; it is not theoretical risk of ruin.

Basic strategy reference

Six decks · S17 · DAS · late surrender · no resplitting · hard double restrictions 9–11. D = double only if eligible.

Hard 9 is explicitly hit vs 2; double vs 3–6; hit vs 7–Ace. Pair 4s split vs 5–6 assumes DAS. All 280 displayed cells have been compared with the documented six-deck S17 reference matrix (280/280 matches). This is a reference-chart check, not independent expected-value certification; rule variants are not dynamically re-solved.

Illustrious 18 and Fab 4 audit reference

Customer-facing index recommendations are disabled pending independent EV and rule validation. The historical table is retained for audit only.

Common six-deck S17 Hi-Lo indices carried forward from your workbook. Apply at or above the listed floored true count.

GroupHandDealerIndexBasic actionDeviation
Some index plays are rule-dependent. The original workbook flags restricted doubling and surrender composition as areas needing independent validation. Treat this as a reference list, not a certified optimal strategy.

Rules implemented

  • Simulator configuration: 1–8 decks, selectable S17/H17, cut-card penetration, bankroll, and table maximum.
  • Dealer soft-17 behavior is configurable; S17 is the default.
  • Blackjack pays 3:2.
  • American hole card with dealer peek on Ace or 10.
  • Double after split is configurable; doubles are restricted to hard 9–11.
  • Late surrender before hitting, after dealer peek.
  • Split-hand limits are configurable; resplit aces and hit-split-aces rules are configurable.
  • Minimum bet $5; starting bankroll and table maximum are configurable. Kelly sensitivity is capped by the selected table maximum.

Model boundaries

  • The chart and indices are transcribed from the supplied workbook, not independently certified. Index deviations are disabled in the advisor; the index table is audit-only.
  • The simulator uses finite-shoe draws, Hi-Lo tags, American hole-card peek logic, configurable split/resplit rules, doubles, late surrender, insurance and S17/H17 dealer play. It does not model ENHC or no-peek round settlement.
  • It does not claim casino-perfect composition-dependent basic strategy.
  • I18/Fab 4 strategies C/D are disabled in the customer-facing simulator pending independent validation.
  • Kelly requires user-supplied edge-at-zero, edge-per-count and return-variance assumptions; the app does not validate them.
  • Observed bankroll-depletion frequency is reported across independent seeds; it is not a theoretical risk-of-ruin estimate. Confidence intervals use a Student-t interval across independent seed/session means.

Responsible use

Blackjack has a house edge under ordinary play. Card counting does not guarantee profit, and short-term outcomes can differ substantially from expected values. Use the app for study and analysis, not as a promise of income. Verify the exact casino rules and legality of play in your jurisdiction.