Every one of these errors feels right at the table. That's exactly why they cost so much. Here's the math, the psychology, and the correct play — for the ten mistakes that quietly drain bankrolls hand after hand.
There are roughly 340 unique starting situations in blackjack — every combination of player total, dealer upcard, and pair situation. Basic strategy tells you exactly what to do in each one. It's been computed, verified, and re-verified for sixty years. And yet, walk into any casino — or watch any free table for fifteen minutes — and you'll see the same handful of mistakes repeated by player after player.
These ten mistakes aren't obscure. They're the common ones, the ones that feel right in the moment but quietly add 1-3% to the house edge. Get rid of them and you'll play closer to the theoretical 0.5% edge that makes blackjack one of the best games in any casino.
This is the most-offered mistake in the casino, and it's offered for a reason: the house wins about 7% of the time you take it. The dealer shows an Ace, and a voice in your ear says "insurance against blackjack?" — and your instinct, especially if you have a strong hand like 20, is to protect it.
The math doesn't care about your strong hand. Insurance is a separate side bet. The dealer needs a 10-value card under that Ace, and there are 16 ten-value cards out of 49 unseen cards in a fresh shoe — roughly 32.7%. Insurance pays 2:1, which requires the dealer to have blackjack 33.3% of the time just to break even. You're conceding nearly 6 percentage points before the cards even resolve.
The only situation where insurance becomes correct is when you're counting cards and the true count is at +3 or higher — meaning the deck is heavily loaded with 10-value cards. If you're not counting (and most players aren't), the answer is always the same word: no. Even when you have a natural 20. Even when the dealer "looks confident." Especially then.
For more on this side bet specifically — including "even money" offers, which are mathematically identical to insurance — see our deep dive on insurance and surrender.
This is the single most-debated decision in blackjack, and the reason is obvious: hitting 16 feels like suicide. The deck has more cards that bust you (any 6 or higher) than cards that help. So you stand, hope the dealer busts, and try to win 23% of the time the dealer goes over 21.
But here's the inconvenient math: the dealer makes 17 or better from a 10 upcard about 77% of the time. Which means standing on 16 vs 10 is essentially conceding 77% of your hands as guaranteed losses, because any dealer score of 17 to 21 beats your 16.
Hitting? You lose 77% of the time too — but you win more often than you'd think, because when you don't bust, you sometimes draw to 17, 18, 19, or 20. The expected loss from hitting is about 53 cents per dollar wagered. From standing, it's about 54 cents. Hitting is marginally better, every single time.
The exception: if you happen to hold 16 made of three or more small cards (like 5+5+6, "many-card 16"), the deck is now slightly more loaded with high cards, and standing is correct. In a single-deck game with proper composition awareness, this matters. In a six-deck shoe, just hit.
If your casino offers late surrender, that's the truly best play here — surrender 16 vs 10 every time. You give up half your bet and walk away. Math approves.
A pair of 8s is a hard 16 — the worst total in blackjack. Splitting them gives you two starting hands of 8 each, which is still mediocre, but at least you're playing two hands that have actual upside instead of one hand stuck at 16.
The fear that prevents splitting is rational: "If I split 8s vs a dealer 10, both hands will probably lose, and I'll have doubled my loss." That's emotionally true and mathematically wrong.
Standing on 16 vs 10 loses you 54 cents per dollar. Splitting 8s vs 10 loses you about 47 cents per dollar — for each split hand. That's 47 + 47 = 94 cents lost across two bets totaling $2, or 47 cents lost per dollar wagered. Splitting saves you about 7 cents on the dollar. Over a year of play, that's a lot.
The principle: always split 8s. Not "split 8s except against 10 or Ace." Always. Even against an Ace, where some old strategy charts hedged. Modern computer-verified basic strategy says split, every time. For the deeper logic on every split decision, see our guide on when to split pairs.
The opposite mistake: splitting when you shouldn't. A pair of 10-value cards is a hard 20 — the second-best hand in blackjack, and the most likely to win. Splitting it turns one near-certain winner into two probable winners, which sounds like math improvement but isn't.
Hard 20 wins about 85% of the time. Each split 10 (starting fresh) wins about 64% of the time. So you'd trade an 85% winner for two 64% chances — yes, you'd win more hands on average, but you'd win less money because you doubled your exposure. Run the numbers: keeping 20 yields about 70 cents profit per dollar; splitting yields about 56 cents per dollar across both bets. You lose 14 cents on the dollar by splitting.
The temptation is strongest when the dealer shows a weak card like 5 or 6. "The dealer is going to bust, so I should have two hands out there!" No. You already have a hand that beats 17 through 20. Don't give that up.
The only exception: serious card counters in deeply positive counts will split 10s in narrow circumstances. Unless you're counting, the rule is simple: never split 10s.
Soft hands — hands with an Ace counted as 11 — confuse a lot of players. A soft 18 (Ace + 7) feels like a strong hand. It's 18, after all. So players stand.
But soft 18 vs a dealer 9, 10, or Ace is actually a losing hand if you stand. Why? Because the dealer is showing a card that frequently makes 19, 20, or 21. Your 18 ties at best, loses often, and never wins outright against those upcards.
The correct play is to hit, because the Ace gives you a free shot. If you draw a small card, your total improves (Ace + 7 + 3 = 21). If you draw a big card, the Ace flips to 1 and you have a hard hand to work with. You can't bust on the next card no matter what comes. The downside is essentially zero, and the upside is meaningful.
This is the most-skipped lesson in basic strategy because the chart shows you should hit a hand that feels strong. Trust the chart. The math says you'll win more often and lose less often by hitting soft 18 vs 9, 10, A than by standing. For complete soft-hand strategy, see when to hit or stand.
A starting hand of 11 is the strongest doubling situation in blackjack. You can't bust on the next card (any card from Ace to 10 keeps you at 21 or below). One in three cards is a 10-value, which makes 21. The dealer's upcard is showing weakness in most cases.
And yet, casual players routinely just hit 11 instead of doubling. Sometimes it's because they don't know the rule. More often, it's because they don't want to put more money in the middle. "What if I get a small card?" you think. "Then I'm stuck doubling on a weak total."
The math says double anyway. The expected return from doubling on 11 versus any dealer card from 2 through 10 is positive — meaning you make money on average. Hitting (with the option to keep hitting) is also profitable, but less so, because you're not maximizing your bet on a great hand.
The one situation where it's debatable: 11 vs dealer Ace. In S17 games (dealer stands on soft 17), basic strategy says hit. In H17 games (dealer hits soft 17), basic strategy says double. Know your table's rule. See our breakdown of when to double down for the full chart.
This isn't a play mistake — it's a table mistake. Increasingly common in Las Vegas and elsewhere, 6:5 blackjack pays only $6 for every $5 wagered on a natural blackjack, instead of the traditional $15 for every $10 (the 3:2 ratio).
It sounds like a minor difference. It's not. The house edge on otherwise-identical 6:5 rules is roughly 1.94%, compared to 0.45% for 3:2. You've quadrupled the house's expected take per hand.
Consider what that looks like: at a $25 table, playing 60 hands per hour, your expected loss at 3:2 is about $7. At 6:5, it's about $29. Per hour. You'd have to be a card counter just to undo the damage from the rule change, and even then it would be marginal.
The trick is that 6:5 tables often look identical to 3:2 tables. The dealer hands cards the same way. The minimum bet might be lower (which is the trap — lower minimums are usually 6:5). Always read the felt before you sit down. If it says "BLACKJACK PAYS 6 to 5," get up and find a different table. We've written a whole article on blackjack payouts if you want the full picture.
The Martingale system says: after every losing hand, double your bet. Eventually you'll win, and your win will recover all previous losses plus your original stake. It sounds bulletproof. It's not.
Two things destroy Martingale in practice. First, table limits. A $10 table typically has a $1,000 or $2,000 maximum. After seven straight losses (10 → 20 → 40 → 80 → 160 → 320 → 640 → 1280), you've already exceeded most table limits. You can't make the bet that would recover your losses, and you've lost $1,270 trying to win $10.
Second, your bankroll. Seven straight losses with Martingale requires $1,270 to keep the streak alive. Sounds extreme — but losing seven hands in a row happens about once every 130 sessions in blackjack. That's not "never." That's a Tuesday.
The fallacy underneath Martingale is the gambler's fallacy: the idea that losses make wins "due." They don't. Cards have no memory. Each new hand is independent of the last. If you've lost five in a row, the probability of losing the sixth is identical to the probability of losing the first.
Use flat betting. Or, if you're varying bets, vary them based on edge (counting) rather than based on emotion (chasing). For a full discussion, see our article on blackjack betting systems — short version: none of them beat the house edge.
This is the gambler's fallacy in a different costume. "This table is hot, I'm staying." "This table is cold, let me find a new one." "The dealer is on a hot streak." "The dealer always wins on this table."
What's actually happening is variance, the mathematical certainty that even fair games produce streaks of wins and losses that look meaningful but aren't. In 100 hands of blackjack, the probability of seeing at least one streak of 6 losses in a row is about 75%. It happens to almost everyone, almost every session.
The illusion of "hot" and "cold" comes from the human brain's pattern-matching, which evolved to detect predators in the bushes — not to evaluate the independence of random card draws. We see streaks and assign them meaning. The cards do not care.
The practical cost: chasing "hot" tables means staying past your stop-loss. Avoiding "cold" tables means table-hopping with no strategic purpose. Both behaviors burn bankroll and time. The cards at the next table behave identically to the cards at this one. There's no escape from variance — only acceptance of it. Our piece on blackjack variance and standard deviation covers what to actually expect over different session lengths.
This is the mistake that bankrupts otherwise-skilled players. You sit down with $500 at a $25 minimum table. Twenty bets deep, if everything goes wrong, you're out. That sounds like a lot of bets. It isn't.
Blackjack has high variance. Over 100 hands, your bankroll can swing by 10-15 betting units (so $250-$375 at a $25 table) just from normal variance, even with perfect basic strategy. Twenty units of cushion means you might be tapped out before the session-level math even has a chance to play out.
The general rule used by professionals: your session bankroll should be at least 50 times your minimum bet, and your trip bankroll should be at least 5x that. So if you want to play $25 tables, bring $1,250 to the session and have $6,250 set aside for the trip. That sounds excessive — it's not. It's what the variance math requires for you to actually survive bad runs and benefit from the long-run edge.
If those numbers are too big, play smaller tables. A $5 minimum game lets you play comfortably with a $250 session bankroll. Lower stakes don't mean less fun — they mean longer play and less stress. Read our bankroll management guide for the full math, including Kelly betting and risk-of-ruin calculations.
The above ten are the heavy hitters. A few more that didn't make the top ten but deserve mention:
Here's the uncomfortable truth: most casual blackjack players are losing 3-4 cents per dollar wagered to the house. Players who play basic strategy correctly lose about 0.5 cents per dollar. That gap — about 3 percentage points — is almost entirely the ten mistakes above.
If you fix the strategy mistakes (1-6), you've recovered most of the gap. If you also fix the psychology and bankroll mistakes (7-10), you've recovered the rest. Card counting takes you the final mile to break-even or slightly positive expected value, but you don't need counting to play well — you just need to stop making these ten mistakes.
Practice them at our free blackjack tables until the right moves are automatic. The free tables are designed exactly for this — make every mistake here so you stop making them where money's at stake.
Long-term, the Martingale fallacy (mistake #8) is the worst because it can wipe out a bankroll in a single bad streak. In terms of cumulative cost over many hands, playing 6:5 tables (mistake #7) is the most damaging — it quadruples the house edge on every single hand.
Against a dealer 7 or higher, yes — hit. Against dealer 2-6, stand (the dealer is more likely to bust). The "always stand on 16" rule is a simplification that costs you money against dealer 7-Ace. If your casino offers surrender, surrendering 16 vs 10 is the optimal play.
It is terrible — both hands will probably lose against an Ace. But the math says splitting loses less than standing on 16. Hard 16 vs Ace loses about 67 cents per dollar. Splitting 8s vs Ace loses about 64 cents per dollar across two bets. It's a damage-reduction play, not a winning play.
Because players keep sitting down. Casinos have found that lower table minimums (often $5 or $10 instead of $25) attract enough recreational players who don't know the rule difference. The cost to the casino of advertising "6 to 5" on the felt is essentially zero, and many players miss it. The lesson: always look.
Avoiding these mistakes brings your house edge down to about 0.5% — the lowest in any major casino game, but still negative. To actually beat the house, you need card counting or some other advantage technique. What avoiding these mistakes gets you is the cheapest entertainment in the casino — a few cents per hand expected loss instead of several cents per hand.
Most players who study with a chart open beside them reach automatic recall in about 500-1,000 hands. That's roughly 8-15 hours of focused play. Trying to learn basic strategy "by experience" without the chart takes orders of magnitude longer and many players never quite finish — they pick up a few rules and approximate the rest. Use the chart.
The dealer gets blackjack about 4.8% of the time (roughly 1 in 21 hands). Over a 100-hand session, the expected number is about 5. Variance means you'll sometimes see 0, sometimes see 10 in a single session. That's normal. If you're playing online at a verifiably fair site, there's no rigging — just human pattern-matching where none exists.