Page 1 of 1

CHoCH on M15, entry on M1: how I keep it from becoming overfit

Posted: Thu Sep 17, 2026 3:47 am
by LondonScalper
CHoCH on M15 plus M1 entry can become a Rorschach test if you let it.

My guardrails: M15 change of character must be obvious without zoom gymnastics; M1 entry must be a standard pattern I already trade — reclaim, pullback hold, retest — not a new shape invented because M15 flipped; and if M15 CHoCH happens into news, I ignore it for scalping purposes until the tape settles.

One change of character per session idea. I do not stack nested CHoCHs across five timeframes into paralysis. Bias from M15, trigger from M1, invalidation written before entry. If any of those three is fuzzy, I pass.

Overfit MTF work feels sophisticated and often trades worse than a simple London ORB with costs counted.

I would rather miss a subtle CHoCH than trade a story that only exists because I wanted a Monday fill.
  • Do you use CHoCH as bias only, or as a trigger?
  • How do you stop MTF structure work from overfitting?
Keen on practical limits people actually follow.

Re: CHoCH on M15, entry on M1: how I keep it from becoming overfit

Posted: Sat Oct 03, 2026 7:40 am
by PropScalpDesk
Bias only. The M1 trigger is one of the same few patterns I'd trade without any M15 context, much like your second guardrail.

What keeps it from turning into a Rorschach test is a mechanical definition. I use 3-bar fractals on M15 for swings (a high with lower highs on both sides), and a CHoCH is a close beyond the latest opposite swing. Two people looking at the chart get the same answer.

The other habit is order. I write the M15 bias and its invalidation with a time stamp before opening M1. Look at M1 first and you start finding M15 structure that agrees with it.

Re: CHoCH on M15, entry on M1: how I keep it from becoming overfit

Posted: Sat Oct 17, 2026 1:08 am
by Shadow Trader
Bias only for me as well, and I would add an expiry to it. If the M15 change of character has not produced one of my standard M1 triggers within eight M15 bars, so two hours, the bias is deleted rather than carried into the afternoon.

Without that expiry I caught myself treating a morning CHoCH as permission at 15:00, long after the structure behind it had been traded through. The time stamped note makes this easy to enforce, because the deadline goes on the same line.

On the fractal definition, I use five bars instead of three, with two lower highs on each side. With three bars on M15 EURUSD I was getting a new swing almost every hour in London, which made nearly any close look like a change of character.

The cost is lag. A five bar swing is only confirmed thirty minutes after it prints, so some CHoCHs appear late enough that the M1 trigger has already gone.

I accept that trade. Fewer, later signals were easier to follow honestly than a steady stream of early ones.

How many CHoCH events per week does the three bar version give you on M15? Mine averaged around four with five bars.

Re: CHoCH on M15, entry on M1: how I keep it from becoming overfit

Posted: Sun Oct 18, 2026 2:24 am
by Shadow Trader
One overfitting check I left out last time is testing the neighbours of whatever setting you picked. After settling on five bar swings, I ran four and six bar versions over the same six months of EURUSD to see whether the result held up on either side of it.

All three landed within a few hundredths of an R of each other, which is what I hoped to see. If five had worked while both neighbours lost money, I would have been looking at a fitted number rather than a real structure rule.

The second check was a pair I had never studied with this method. GBPUSD gave fewer signals because its swings are wider, but the M1 triggers behaved much the same, so I kept the rule unchanged instead of tuning it for the new pair.

There is one edge case I now filter by hand. When the bar that closes beyond the swing is the 07:00 London candle itself, the break is usually just the opening spike, and I wait for the next M15 close to confirm it.

Do you treat the open candle any differently, PropScalpDesk? With three bar fractals I would expect that first bar to produce plenty of false breaks.

Re: CHoCH on M15, entry on M1: how I keep it from becoming overfit

Posted: Sat Oct 31, 2026 4:27 am
by Shadow Trader
The setup appears. The rules say enter. The hand doesn't move. Half a second passes, then a second. The price has gone.

Hesitation at entry is common among traders who have had a string of losses. The brain wants to confirm one more time, and the confirmation takes just long enough to ruin the entry.

Often the hesitation is a signal. It might mean the setup isn't truly clear, or that the trader doesn't trust the system at that moment. Forcing the entry in such a state usually doesn't help.

Other times it's just fear, and the setup is perfectly valid. Distinguishing the two requires honest review. Look at the trades skipped through hesitation and compare their results with those taken without hesitation.

If the skipped ones perform just as well, the hesitation is costing money. If they perform worse, it's useful intuition.

Practical measures include reducing size to a level where the fear shrinks, using pre-placed orders so the decision is made earlier, and writing out exact entry criteria so there's less to evaluate in the moment.

Which kind of hesitation shows up more often: the useful kind, or the costly one?