There is one principal mistake: reading a divergence as a reversal signal. Almost all the others follow from it.
Divergence as a signal
A mismatch describes what already happened: the current move's impulse is weaker than the previous one by the chosen indicator's measure. Neither a reversal nor its timing follows from that.
A divergence can persist for months, renewing at every new extreme while the trend carries calmly on. That is not a "false divergence" — it is genuine, it simply does not mean what people expect of it.
Hunting on an unconfirmed extreme
A mismatch is built on two neighbouring extremes. At the fresh edge of the chart the later one is not yet confirmed: price may go further, and what looked like a peak turns out to be the middle of a move.
A divergence on the last bar is the least reliable of all, and it is the one people notice most.
Stacking mismatches as confirmations
"Divergence on RSI, on MACD and on volume — triple confirmation." No: all three are computed from the same price series. They are not independent, and agreement more often means one observation counted three times.
Disagreement between indicators is more informative than agreement: it shows the conclusion depends on which measure you picked.
Tuning settings until a mismatch appears
Divergence depends entirely on what counts as an extreme. Relax the requirement on reversal size and new peaks appear, and with them new mismatches.
Turn that dial until a divergence shows up and one always will. Settings are worth fixing before you look at the result.
Reading the indicator's curve rather than the bars
The indicator's value is taken in the same bar as the price extreme — with a small tolerance, because a smoothed indicator is usually slightly offset.
Comparing two curves by eye without tying them to specific bars produces mismatches where there are none: lines on different scales can be tilted to suit.
Confusing regular with hidden
A pair of extremes holds exactly one mismatch or none: the price conditions are mutually exclusive. If both appear in the same argument, something is wrong in the definitions rather than doubly confirmed.