Wait for the Candle Close or Enter on the Live Arrow? Early Entry Won 7 of 10 on M15, Waiting Won 8 of 10 on H1

Every indicator forum gives the same advice: never act on an arrow until the candle closes, because it may disappear. That is true, I measured it. In the non-repaint indicators re-check, 32.5% of arrows seen during a bar were gone by its close. But waiting has a price too: you enter later, at a worse price, on the arrows that were real. So which costs more, the ghosts or the delay?

The test: EMA 5/20 cross, MACD 12/26/9 signal cross, RSI 14 back out of 30/70, Stochastic 14/3/3 cross below 20 or above 80, and Supertrend 10/3 flip. EURUSD and gold, June 3 to August 26, 2026. M15 and H1 bars built from M5, with a look at the forming bar every 5 minutes. Early entry: buy or sell at the first snapshot where the arrow appears, even if it vanishes later. Close entry: buy or sell at the bar close, only on arrows that are still there. Both exit 2 hours after entry. Cost 1 pip on EURUSD and $0.30 an ounce on gold.

Which entry won on each chart?

Average per trade after cost. Gold in dollars per ounce. “Early” includes every arrow that appeared during the bar, whether it survived or not.

On M15 the early entry won 7 times, on H1 it won 2 times. Gold H1 went to the close 5 out of 5. The rule “wait for the close” is right for the hourly chart and mostly wrong for the 15-minute one, at least on these weeks.

Why does the timeframe flip the answer?

Two things get worse at the same time on H1. More arrows vanish, 39% instead of 30%, because an hour gives price time to cross back. And each ghost costs more, because the swing that cancels an hourly arrow is bigger: ghost trades on gold H1 lost $6.16 to $16.48 an ounce, on gold M15 $1.47 to $7.40. On M15 the wait is short and the ghosts are cheaper, so getting in a few minutes earlier on the real arrows can be worth it.

What if you could tell real arrows from ghosts?

That is the whole trade-off in one table. An early entry on a real arrow beat the close entry every time: on EURUSD H1 RSI it was +8.10 pips a trade against +3.62. An early entry on a ghost lost every time, and beat the close entry once, on EURUSD H1 Stochastic, where both lost. But in the moment you cannot tell which one you are looking at, so the early trader gets both, and the mix decides. The same lesson about how an arrow looks in hindsight and how it trades live came out of the repainting arrows test, where fractals looked great in history and lost when traded two bars later.

Does either way make these indicators worth trading?

Not on their own. Entering at the close made money in 9 of 20 match-ups and in both halves of the period in 7. Entering early made money in 7 and in both halves in 4. The positive ones were mostly on gold, which trended over these weeks, the same pattern as in gold vs EURUSD with one rule. On EURUSD M15 nothing made money after the spread either way, which matches the moving average crossover test and the Stochastic signals test.

So should you wait for the candle close?

• On H1 and above, yes. Waiting won 8 of 10 match-ups, and ghosts on the hourly chart cost the most.
• On M15 the rule did not hold here. Early entry won 7 of 10. If you trade M15 arrows early, accept that about 3 in 10 will vanish and lose.
• Back-tests describe close entries only. An MT5 indicator back-test uses the closed bar, so an early-entry trader is trading something that was never tested. How to check if an MT5 indicator repaints shows how to see bar 0 yourself in the Strategy Tester.
• Fix the entry before you fix the timing. Most rules here lost either way. A filter or an exit rule matters more, see trading system vs entry signal.

Frequently asked questions

Should I wait for the candle to close before entering?

On H1 waiting won 8 of 10 match-ups in this test. On M15 entering on the first arrow won 7 of 10. Overall 11 to 9 for waiting.

How many indicator signals disappear before the close?

32.5% of arrows seen during the bar across five indicators: about 30% on M15 and 39% on H1.

Do ghost signals lose money?

Yes, in all 20 match-ups. They appear and vanish because price turned against them, so the early entry starts on the wrong side.

Is entering at bar open of the next candle the same as entering at the close?

Almost. On MT5 the next bar opens at or very near the last close, so the close entry here is what an EA reading the closed bar would get.

Which timeframe is safest for arrow indicators?

None was safe, but the hourly chart punished early entries most and rewarded waiting most.