⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.
When SCGP failed, I spent time looking at the chart for what I'd missed. I found nothing. The setup was clean. The base was tight. The volume expanded on the break. By every rule I'd built, the trade was correct.
The problem was never on the chart. The SET index fell 7.4% in the two weeks after that breakout. Another stock — STA — had an equally clean breakout the same week and collapsed the same way. Two unrelated setups. Same failure. Same timing. The market had turned, and it took everything down with it.
That experience points at something the data confirms very clearly: the market condition is not a footnote to your setup. It is most of the outcome.
What the numbers say
I split my 36-year Thai backtest by market condition — the same setup (RS≥80 leaders, contracting base, pivot breakout) across different market environments. The full results are in [The Market Regime Gate](/articles/market-regime-gate.html).
The setup doesn't produce edge in a vacuum. The edge exists almost entirely inside confirmed uptrend windows. Outside those windows, you're not trading a system with an edge — you're trading a system with a coin flip dressed up as pattern recognition.
The US data says the same thing, even more sharply. Without a market uptrend filter, the US breakout edge is −0.02R — statistically indistinguishable from zero. With RS≥80 and confirmed uptrend: +0.12R. The filter doesn't improve the system. The filter is the system.
What a confirmed uptrend means (and why an MA cross misses it)
Before I built a proper market condition tracker, I used a moving-average cross as a regime filter. The 50-day crossing the 200-day — the "golden cross." It felt like the obvious approach.
It's blind.
A moving-average cross lags by months. By the time the 50d crosses the 200d, a bull market is often already mature. By the time they cross back, a bear market is often well underway. You're not reading the market's current health — you're reading where the market was three months ago.
The method I use now is adapted from IBD's distribution-day count: track days when the market closes down significantly on higher volume than the prior session — evidence of institutional selling. When those count up (typically 5–6 within a rolling window), the market is "under pressure" or "in correction." When they're absent and the market is in a follow-through day or price uptrend, it's a confirmed uptrend. When distribution clears or a follow-through fires, it resets.
This is a current read, not a lagged one. It tracks what institutions are doing right now, not what happened over the past hundred days.
The Thai regime finding: the signal reverses
The regime sensitivity isn't just about being in or out of the market. In Thailand, I tested a specific signal — a character-change pattern on the RS line — across different market conditions.
The result was stark:
- Regime UP (SET above both 50d and 200d, 50d rising): median forward return +2.95pp above baseline, CI excludes 0 — a real signal.
- Regime DOWN: the same signal flips to −1.48pp — a negative result.
The signal doesn't just weaken in bad conditions. It inverts. The same pattern that predicts outperformance in an uptrend predicts underperformance in a downtrend.
This is why treating market condition as a soft preference ("I prefer to trade in uptrends") is not the same as treating it as a hard gate. If the signal actively reverses, trading it in the wrong regime isn't a lower-probability version of the same trade — it's a different trade with the opposite expected outcome.
Why breakout failures are not breakout problems
The SCGP and STA lesson generalizes. In a falling market, breakouts fail at a rate that makes the setup net-negative. This isn't a chart-reading problem. It isn't a setup-quality problem. The individual trade looked fine; the macro context made it a low-probability bet before the chart was even opened.
This means the skill of reading a good setup is not separable from the skill of knowing when not to use it. The first without the second produces breakeven results at best, losses at worst. Together they produce the edge the backtest actually found.
The market gate is not a refinement to the method. It is the frame that makes the method work at all.
The practical implication
Check market condition before you open a single chart.
If the market is in correction, the answer to "should I buy this breakout?" is no — not because the chart is bad, but because the backtest says the expected value of that trade is near zero or negative, regardless of chart quality.
If the market is in confirmed uptrend, open the charts and apply your method.
This sounds obvious. It is obvious. The failure mode is not misunderstanding it — the failure mode is doing the analysis in the wrong order: reading charts first, then looking for reasons the market "isn't that bad." Read the market first. Let the market decide whether you're allowed to play.
Track. Study. Wait — this is what wait means. Strike.
Personal research and trading journal — not investment advice. The author does not provide licensed advisory services. — MOEasymmetry