Imagine two players sitting at the same blackjack table over the same 200 hands. Both win 48% of those hands. Both end the session down roughly the same dollar amount. By every summary statistic, they played identically. But one of them was forced to rebuy midway through. The other was never in danger. The difference between them wasn't skill, luck, or strategy. It was sequence.
This is path dependency, and it is one of the least discussed forces in practical gambling. The order in which outcomes arrive changes what happens to your bankroll in ways that the final win rate cannot capture—and once you understand the mechanism, a number of common gambling habits look considerably less reasonable.
What Path Dependency Actually Means
In mathematics, a process is path-dependent when the sequence of events, not just their sum, determines the outcome. Gambling bankrolls are path-dependent by construction.
Here's a clean example. Suppose you start with $1,000 and play a game where you bet $100 per hand. Scenario A: you lose the first five hands, then win the next five. Scenario B: you win the first five, then lose the next five. Both sequences produce the same net result—zero profit, zero loss, back to $1,000.
Except Scenario A took your bankroll down to $500 before the recovery began. At that low point, you were at risk of hitting your stop-loss, running out of funds, or—critically—making different decisions than you would have made with $1,000 in front of you. Scenario B never put you in that position. Same ten hands. Same outcomes. Completely different experience of risk.
This is not a minor nuance. The low point your bankroll reaches during a session is as important as where it ends up, and that low point is entirely a function of sequence.
The Bet-Sizing Amplifier
Path dependency becomes substantially more damaging when players adjust bet sizes during a session—which most players do, in one direction or another, even when they intend not to.
Consider what happens with proportional or variable betting. If a player increases bets after wins (pressing) and decreases after losses (pulling back), then the sequence of outcomes determines which bet sizes are attached to wins and which are attached to losses. Win big on small bets, lose big on large bets—that is the hidden cost of being in a losing sequence while pressing.
The reverse framing makes the point more sharply. A player who catches their hot streak early—when bets are large—and absorbs their losses later—when they've scaled back—extracts more money from the same win rate than a player whose sequence runs the opposite direction. Identical underlying results, meaningfully different dollar outcomes.
Flat betting reduces but does not eliminate path dependency. Even with fixed bet sizes, a losing sequence early enough and deep enough can compromise your ability to continue playing at all—which means the sequence controls whether the law of large numbers has enough hands to work with.
Ruin Before Recovery
The concept that formalizes this most usefully is ruin before recovery. Your bankroll does not care about your long-run expected results if a specific sequence depletes it before the long run arrives.
This is why starting bankroll interacts with sequence in a way most players don't fully account for. A deeper bankroll doesn't change your expected value. It doesn't improve the game's odds. What it does is reduce the probability that an early adverse sequence eliminates you from the session before variance can smooth out. It buys you more paths—more possible sequences—and some of those paths lead to recovery that a shallower bankroll would never reach.
The practical implication: when you're calculating how much to bring to a session, you aren't just calculating average expected loss. You're calculating how deep your worst plausible sequence might go before mean reversion begins pulling you back. Those are different numbers, and sessions are frequently sized for the former when they should be sized for the latter.
Reading Your Results More Accurately
Path dependency also has consequences for how you interpret past sessions. If you ran well early and coasted through a session, your logged win rate may reflect a sequence that was favorable to you in ways that won't repeat consistently. If you clawed back from a large deficit, your result understates how close to ruin you actually were.
This matters when players use session histories to estimate their skill level or calibrate future strategy. A player with ten winning sessions might have benefited from favorable sequences in seven of them—not better play, just outcomes arriving in a friendlier order. The win rate looks identical either way.
Correcting for this is difficult at the individual session level, which is exactly why sufficiently large sample sizes matter. Over enough sessions, the sequence effect averages out. But most players make strategic decisions—about bet sizing, game selection, session length—on sample sizes where sequence noise still dominates the signal.
What You Can Actually Do With This
A few things become clearer once you take sequence effects seriously.
Fund sessions for the path, not the average. Estimate the realistic worst-case sequence for the game and stake you're playing—not just expected loss—and bring enough to survive it without making distorted decisions. A rough rule: if your game's variance is high, your session bankroll should be meaningfully larger than expected loss calculations alone suggest.
Treat early losses as sequence information, not just results. An early adverse run doesn't tell you your edge has changed. It tells you you're in an unfavorable sequence, which may or may not recover. What it definitively tells you is that your remaining bankroll is smaller, so the bet sizes that were appropriate at session start may not be appropriate now.
Be skeptical of session streaks in both directions. A winning streak early in a session often reflects nothing more than a favorable sequence. It is not evidence of an upgraded edge, and sizing up in response converts a random sequence advantage into a structural risk. The same logic applies in reverse—a losing streak is not evidence that your strategy is broken.
Log drawdowns, not just outcomes. If you track gambling results, the most useful addition to a win/loss log is the maximum intra-session deficit. That number tells you more about the sequences you're actually encountering—and whether your bankroll structure is adequate for them—than a final result ever will.
The aggregate statistics you walk away with after a session are real. They're just incomplete. The sequence that generated them is part of the result, and ignoring it is how players misread what happened to them and miscalibrate for what comes next.