EURUSD pushes through 1.0800 support, trades to 1.0788, and every trader watching has the same half second of doubt: is this the real thing, or is it about to snap back through the level and trap everyone who just sold. Ninety seconds later it matters enormously which one happened, and the difference was visible the whole time in three things: whether the candle closed beyond the level, whether the next candle followed through, and how long price actually spent on the new side before anyone started arguing about it.
Acceptance versus rejection
A break is acceptance when the market is genuinely willing to trade at the new price and stay there. It is rejection, a false break, when price only visited the new side briefly before the crowd that defended the old level reasserted control. The tell is not the initial move, which looks identical in both cases for the first few seconds. The tell is what happens in the minute or two afterward: acceptance shows follow-through, more candles closing on the new side, minimal effort to return; rejection shows an immediate stall, a wick straight back through the level, and often a stronger move in the opposite direction than the original break itself, because the trapped breakout traders are now exiting in a hurry.
The close is the first real filter
The single most useful discipline here is waiting for a full candle close beyond the level on your working timeframe before treating a break as real. An intrabar touch, price wicking through 1.0800 and printing 1.0788 before the candle finishes, tells you nothing on its own; plenty of candles wick through a level and close right back above it. A close at 1.0791, clearly beyond the level and holding there through the end of the candle, is meaningfully more informative. This single filter, waiting for the close, eliminates a large share of the false signals that catch traders who react to every intrabar poke.
Follow-through and time as confirmation
A close beyond the level is necessary but not sufficient. The next one or two candles should continue in the direction of the break, or at minimum hold the new side without threatening to return. On Nasdaq, a break of 19,300 support followed by a close at 19,270 that is immediately followed by a bounce back to 19,310 is not showing follow-through, and treating it as a confirmed break invites exactly the trap the false break's reputation warns about. Time matters too: a genuine break tends to hold for at least several candles on the working timeframe, while a false one often reverts within one or two. Neither follow-through nor time alone is proof, but together they separate acceptance from rejection more reliably than the initial move ever will.
What to do with each read
A confirmed break, close beyond the level with follow-through, opens the retest trade described elsewhere: wait for price to return to the broken level and react before entering in the direction of the break. A failed break, the level holding after the poke through, is itself a tradeable event, often one of the sharper reversals available, because it traps everyone who acted on the initial move. The mistake to avoid is treating every break the same way regardless of which read the close and follow-through actually support; the two scenarios call for opposite trades, and confusing them is how a level swallowing a break turns into a loss instead of the honest signal it actually was.