Odds TrainerBlackjack Strategy
Guide · The honest answer

How to Win at Blackjack

By Spencer S. · Updated September 3, 2026

The honest answer

You can't erase the house edge with strategy alone — but you can shrink it from the 2–4% a guessing player loses down to about 0.5%, the best odds in the casino. Do it with four levers: perfect basic strategy, the right table, no sucker bets, and bankroll discipline. Only card counting actually flips the edge to the player.

Search "how to win at blackjack" and you'll drown in systems, secrets, and nonsense. Here's the version with the math attached. Blackjack is the most beatable game on the floor, but "beatable" has a precise meaning — and most of what costs players money is fixable in an afternoon. This is what actually moves your odds, ranked by how much each is worth.

Can you actually win at blackjack?

Two different questions hide inside that one. Tonight? Absolutely — a perfect player still wins about 42.4% of hands, loses 49.1%, and pushes 8.5%, and short-run variance means walking away up a few hundred dollars is completely normal. Forever? No — with a ~0.5% house edge, enough hands grind your bankroll down. The goal of smart play isn't to defy that math; it's to make the edge as small as possible and give variance room to land in your favor. Here's how each lever does that.

The four levers, ranked by impact

LeverWhat it's worthEffort
1. Perfect basic strategy~2–3.5%One chart to learn
2. The right table (3:2, S17)~1.6%Choose where you sit
3. Refuse sucker bets~0.2–8%*Just say no
4. Bankroll & variance controlSurvival, not edgeDiscipline
Card counting (the only true edge)+0.5 to +1.5%Months of practice

*Sucker-bet cost depends which you avoid — insurance is a ~7.7¢ loss per dollar, some side bets carry house edges above 8%. Percentages are house-edge reductions; they don't simply add, but together they take a typical guessing player from a 2–4% hole to about 0.5%.

Lever 1: Play perfect basic strategy

This is the whole ballgame. A player who guesses gives up 2–4% per hand; perfect basic strategy drops that to about 0.5%. That's not a small tweak — it's the difference between the worst and the best odds in the casino, and it costs you exactly one chart to learn. The big leaks, in order:

You don't memorize the chart by staring at it — you drill it. Our free trainer deals real hands and shows the live cost of every wrong button, so the corrections stick.

The most expensive mistakes, priced by our engine

"2–4%" is an average over every hand. The money actually leaks in a handful of specific spots, and they are not the ones people argue about. Every figure below is our engine's modeled value for a fresh 6-deck, dealer-hits-soft-17 shoe, in cents per $1 of your original bet; the cost is the gap between the play people make and the play the chart makes.

The mistakeWrong playRight playCost per $1
Splitting 5,5 vs 6 (it's a hard 10 — double it)+16.8¢+58.7¢41.9¢
Hitting 11 vs 6 instead of doubling+33.9¢+67.8¢33.9¢
Hitting A,A vs 10 instead of splitting−6.6¢+18.4¢25.0¢
Standing on A,7 vs 6 instead of doubling+22.3¢+35.9¢13.6¢
Hitting A,6 vs 6 instead of doubling+12.7¢+25.5¢12.7¢
Splitting 10,10 vs 6 (stand on your 20)+55.5¢+67.7¢12.2¢
Hitting 8,8 vs 9 instead of splitting−50.6¢−40.8¢9.8¢
Standing on A,7 vs 9 (soft 18 isn't "good" here)−18.3¢−9.9¢8.4¢
Standing on 16 vs 7−47.7¢−40.9¢6.8¢
Standing on 16 vs 10 — the famous one−54.2¢−53.5¢0.7¢

Notice the shape of that list. The hand everyone debates, 16 against a ten, is the cheapest mistake on it. The expensive ones are all skipped doubles and mishandled pairs — the spots where you had the edge and didn't press it. If you fix nothing else, fix your doubles and splits. Split values use the engine's approximate split tree, so the ranking is solid and the last decimal isn't; method and caveats are on the methodology page, and the full grid is the expected value chart.

Lever 2: Pick the right table

Two players using identical perfect strategy can face wildly different odds depending on the felt. Before you sit down, check:

What each rule is worth, on the same additive model our house edge calculator uses (six decks, H17, double after split, 3:2 — a 0.62% baseline):

RuleHouse edge change
Blackjack pays 6:5 instead of 3:2+1.39%
Dealer hits soft 17 instead of standing+0.20%
No double after split+0.14%
8 decks instead of 6+0.02%
2 decks instead of 6−0.17%
1 deck instead of 6−0.46%
Late surrender offered−0.08%

Read it top-down: one 6:5 line outweighs every other rule on the list combined. Want to price your exact table? The house edge calculator turns any rule combination into a real number, including the per-hour cost.

Lever 3: Refuse the sucker bets

The casino offers several "extra" wagers that exist only because players take them. Decline every one:

Lever 4: Manage bankroll and variance

This one doesn't lower the house edge — it keeps you alive long enough for the small edge and good variance to matter. Bring a session bankroll you can lose without flinching, bet a small fraction of it per hand, and set a walk-away line in both directions. What does not work is any betting system: the Martingale and its cousins feel clever but just convert many small wins into rare catastrophic losses. No bet-sizing scheme changes the odds of the next hand — here's why.

What your play costs per hour

Expected loss is just bet × hands per hour × house edge. Put in your own numbers and see what perfect play is worth against how most people actually play.

House edge as you play1.62%
Expected cost per hour$28
Per 4-hour session$113
Perfect play, 3:2 table$11/hr

Baseline 0.62% (six decks, H17, double after split, 3:2), 6:5 from the additive rule model, insurance at about 0.3% per hand if always taken. "A few leaks" (+1.0%) and "mostly guessing" (+2.5%) are estimates, not measurements — the real number depends on which mistakes you make, and the table above prices the common ones.

The only way to truly beat the game

Everything above shrinks the house edge; nothing above removes it. The one legal method that hands the advantage to the player is card counting — tracking the ratio of high to low cards and betting more when the deck favors you. It genuinely works (a good Hi-Lo counter plays at roughly a +0.5% to +1.5% edge), it's not illegal, and it's also hard, slow to learn, and unwelcome at the casino. For the 99% who won't count, "winning at blackjack" means playing the smallest possible house edge and quitting while you're ahead. That's not a consolation prize — it's the best deal in the casino.

The fastest way to a 0.5% edge.

Drill perfect basic strategy in the free trainer — it shows the live win chance and dollar cost of every Hit, Stand, Double, and Split, and logs the mistakes that are costing you the most.

Play the trainer free

Frequently asked questions

Can you actually win at blackjack?

In a session, yes — a perfect player wins about 42% of hands and often finishes ahead. Long-term, strategy alone can't beat the ~0.5% house edge; only card counting flips the advantage to the player (about +0.5% to +1.5% in good conditions).

Is blackjack luck or skill?

Both. Skill sets your house edge — perfect play means ~0.5%, guessing means 2–4%. Luck (variance) decides any single session. Skill controls the long-run floor; luck controls tonight.

Can you win without counting cards?

You can win sessions but not the long run. Perfect basic strategy at a 3:2 table is the best you can do without counting — about 0.5% house edge, still negative.

How much can you win at blackjack?

In a session, as much as variance allows — winning runs are real. Sustainably, only a card counter holds an edge, and even then it's a small percentage of total money wagered, earned over many hours.

Keep learning