Indicator Deep-Dive: Not All Markets Are Created Equal — Our Signals Prove It

The Question Every Trader Should Be Asking Their Indicators

Most retail traders treat technical indicators as universal tools — apply RSI to crude oil the same way you apply it to EUR/USD, and expect comparable results. Our proprietary backtest and live signal data, covering 30 closed signals across three asset classes, suggests that assumption is not just wrong — it's quietly expensive. The central question this report addresses is simple: do our indicators perform differently across forex, equities, and commodities? The answer, even with a limited sample, is yes. Materially and directionally so. Understanding where a signal framework earns its edge — and where it destroys it — is the difference between a system you can actually deploy and one that bleeds you slowly across asset classes you thought you understood.

Thirty Signals, Three Markets, One Uncomfortable Average

Start with the headline number, because it deserves scrutiny. Across all 30 closed signals, the overall win rate is 33.3% with an average trade return of -0.74%. That is a losing system at the aggregate level. A 33% win rate is survivable — some trend-following strategies operate profitably well below 40% wins — but only if the average winner is meaningfully larger than the average loser. Our aggregate data does not yet allow us to calculate that ratio with confidence, but a negative average return at this win rate tells us the math is currently not working in our favor. We publish this plainly, because the alternative — selectively reporting only the asset class that looked best — would make this newsletter useless.

What the aggregate conceals, however, is a performance gradient across the three categories that is both pronounced and analytically meaningful. Forex, stocks, and commodities did not behave like the same market running through the same indicator set. They behaved like three different problems.

The Hierarchy Is Stark: Forex Leads, Commodities Lag Badly

Forex is the relative standout. With 16 signals — the largest sub-sample we have — forex produced a win rate of 43.75% and an average return of -0.04%. That near-zero average return on a sub-40% win rate is not a success story, but it is operationally different from the other two categories. At -0.04%, the forex signals are essentially breaking even on average. The system is not generating alpha here, but it is also not destroying capital in a systematic way. For a signal framework still in calibration, that is a meaningful distinction. Sixteen observations is also the only sub-sample that approaches a threshold where patterns begin to have some descriptive validity — though we will return to the caveats on that shortly.

Equities tell a harder story. Eight signals, a 25.0% win rate, and an average return of -1.05%. The win rate drop from forex to stocks — nearly 19 percentage points — is large enough to be directionally informative even at n=8. The average loss per signal in stocks is roughly 26 times worse than in forex on a per-trade basis. That is not noise. Something structural is different about how our current indicator configuration interacts with equity price behavior versus currency pairs. Whether that is a function of mean-reversion dynamics, earnings-driven gap risk, lower liquidity in the specific names we tracked, or simply parameter mis-fit is something the current data cannot answer definitively.

Commodities are the most alarming category, and also the one where we must be most cautious about drawing firm conclusions. Six signals. A win rate of 16.67% — meaning exactly one signal in six closed profitably. An average return of -2.19% per trade. If forex is the patient in observation, commodities is the patient in triage. The average losing trade in this category appears to be pulling the mean down severely, suggesting either that our stop parameters are too wide for commodity volatility profiles, that the indicator signals are systematically late relative to commodity-specific news flow — energy inventory reports, agricultural supply shocks, metals positioning data — or that this sample, small as it is, happened to capture a particularly adverse sequence. We genuinely do not know which. Six signals cannot tell us.

What This Data Does Not Support

Intellectual honesty requires naming the interpretations this dataset cannot sustain, even if the directional story looks clean.

First, we cannot claim statistical significance for any of these sub-samples. With n=16 for forex, n=8 for stocks, and n=6 for commodities, none of these groups meets a threshold for robust inference. The patterns are suggestive. They are worth tracking and acting on at a portfolio-management level — specifically, by reducing exposure in the categories performing worst while we gather more data. But we would be misleading you if we presented these numbers as proof that our indicators categorically fail in commodities and categorically work in forex. They don't yet prove that. They hint at it.

Second, the data does not tell us whether the problem is the indicator, the parameterization, or the signal generation criteria. A momentum indicator miscalibrated for commodity volatility will look broken even if momentum is a valid framework for that market. We are measuring the output of a specific configuration, not the validity of an entire indicator class. That distinction matters enormously when deciding how to respond to this data — whether you recalibrate, or abandon the approach entirely.

Third, we don't know the timeframe distribution of these 30 signals. If the commodities signals were concentrated in a period of unusual low-trend, choppy price action — which commodity markets do exhibit cyclically — the poor results may reflect regime misfortune as much as structural indicator failure. This is preliminary data. We are treating it as directional evidence, not verdict.

Fourth, the overall -0.74% average return does not account for position sizing, compounding, or transaction costs in a live account context. The raw signal performance reported here is the cleanest read we have, but it is not a complete picture of what a trader following these signals would have experienced in a real book.

What Traders Should Actually Do With This

Given the data and its limitations, here are the practical implications we would act on today — not as rigid rules, but as calibration inputs.

• Treat forex signals as the primary deployment zone for now. A -0.04% average return across 16 signals is not a green light to lever up. But it is the closest thing to a functioning signal in our current dataset. If you are going to use these indicators actively, forex is where the damage is most contained and the win rate is closest to actionable territory. Monitor whether the win rate sustains above 40% as the sample grows — that is the level where favorable risk-reward ratios can begin to generate positive expectancy.
• Reduce position size in equity signals, or wait for the sample to grow. Eight signals at 25% win rate and -1.05% average return is a clear underperformance signal relative to forex. We are not saying stop using indicators on stocks. We are saying do not trade equities signals at the same size you would trade forex signals until we understand whether this is a parameterization issue or a structural one. Half-size or quarter-size while we gather data is a reasonable posture.
• Apply strict review before acting on any commodity signal. Six signals, one winner, -2.19% average return. Until we have 20 or more commodity signals closed, we would recommend paper-trading or minimum-size live testing only. The risk-reward picture is too unfavorable and the sample too small to justify standard position sizing. If you must trade commodity signals, ensure your stop-loss parameters are tighter than your defaults — commodity volatility may be consuming stops that would be appropriate in forex.
• Track the next 20 signals across all three categories with obsessive precision. The data will get meaningfully more useful when forex crosses 30 total signals and stocks approaches 20. At those levels, we can begin making more defensible claims about whether these patterns are persistent or artifacts of a short, unrepresentative sample window.
• Ask whether your indicator parameters were optimized on forex data. Many widely-used technical indicators — RSI period defaults, moving average lengths, Bollinger Band widths — were developed and popularized in the context of equity or forex markets. If our signal system was tuned primarily on currency pair behavior, it would not be surprising to see it underperform in the different volatility structures of commodity markets. This is a question for our research team to investigate in the next cycle.

A Note on Our Methodology and Its Honest Limits

This report is based on 30 total closed signals generated by Stocks365's proprietary backtest and live signal system, spanning three asset classes: forex (n=16), stocks (n=8), and commodities (n=6). We have not disclosed the specific timeframe over which these signals were generated, the specific instruments or currency pairs included, the holding period per signal, or the exact indicator combination being evaluated — partly for competitive reasons, partly because this is an early-stage data set where those specifics are still being standardized across our signal pipeline. The numbers reported here represent raw per-trade returns, not risk-adjusted or compounded performance figures. No claim of statistical significance is made for any sub-sample. We are publishing this data precisely because the results are unflattering in many respects — a 33.3% overall win rate and negative average return are not outcomes we want to hide behind qualifications. The honest purpose of this report is to document where we are, not where we want to be. As we close additional signals, we will update these figures publicly.