The gambler fallacy: nothing is "due"
The gambler fallacy is believing that past independent results change future ones — that after five straight heads a coin is "due" to land tails. It is not. Each flip is still 50/50, because the coin has no memory.
Games are close to independent in the same way. A team that has failed to cover four weeks running is not "due" to cover, and a total that has gone over five times is not "due" to go under. Treat each bet on its own merits and price, not on a streak that feels like it has to end.
Chasing losses is how bankrolls die
Chasing is trying to win back a loss by betting bigger — often doubling up after each loss (the "martingale"). It feels logical and it is ruinous. Suppose you try to recover a $100 loss by doubling: $100, then $200, then $400, then $800. A four-bet losing streak is not rare — at a 55% win side it happens about 4% of the time — and by then you have risked $1,500 to try to win back the original $100.
One bad run wipes out the account. The fix is mechanical: fixed unit sizes, set before the day starts, that do not move because you are stuck. Your stake should reflect your bankroll and edge, never your mood or your losses.
Recency bias and the sunk-cost trap
Recency bias is overweighting what just happened — riding a "hot" team or fading a "cold" one based on a tiny recent sample. As the sample-size lesson showed, a few games carry almost no signal. Anchor on the full body of evidence and the price, not last week.
The sunk-cost trap is throwing good money after bad because you have already invested — staying on a losing model or a live bet just to avoid admitting the first stake was a mistake. Money already wagered is gone; the only question that matters is whether the next bet has value on its own. Judge every bet forward, never backward.