Most indicators don't fail because the maths is wrong. They fail because of how they were built, how they were shown to you, and how you were always going to use them.
Four failure modes cover nearly all of it.
1. Lag is not a bug you can settle
Every indicator derived from price is a transformation of data that has already happened. A 20-period moving average cannot tell you anything until 20 periods have traded. An oscillator cannot register momentum until momentum exists.
The usual reaction is to shorten the lookback — faster settings, earlier signals. That doesn't remove the lag; it lowers the threshold at which noise qualifies as signal. You trade fewer late entries for more false ones.
There is no setting that escapes this. There is only a decision about which error you'd rather make, and the decision should follow your holding period:
Short holds cannot absorb a late entry, so they need faster inputs and a filter to reject the extra noise.
Longer holds can absorb lateness easily, so slower inputs and fewer signals cost almost nothing.An indicator that doesn't acknowledge this trade-off — that markets itself as "no lag" — is either shifting plots backwards in time or firing on the unclosed candle. Both are covered in the repainting guide.
2. Curve-fitting: the settings were chosen after the fact
Give anyone a chart and enough parameters and they will find a combination that looks extraordinary on that chart. Length 14 didn't work, so try 17. Try 21 with a 3-bar smoothing and only in the London session. Eventually the historical picture is beautiful.
What has been discovered is the noise in one specific slice of history, not a property of the market. Forward data has different noise, and the fit dissolves.
Tells that a tool is fitted rather than designed:
Oddly specific defaults. A 14-period lookback is convention; a 17-period lookback with a 2.3 multiplier is a search result.
A lot of parameters. Every additional knob multiplies the number of combinations that can be tried, and one of them will always look great.
Examples all drawn from one market and one period. If a tool only ever demonstrates itself on 2021 crypto, ask what it did in 2022.
Settings that must be "optimised per instrument." Sometimes legitimate for volatility scaling. Often a way of pushing the fitting exercise onto you.The defensible alternative is simple, conventional inputs that scale with volatility rather than being hand-picked per chart, and an explanation of why each one is there.
Stacking indicators feels like building a case. Usually it's asking one question five times.
RSI, Stochastic, MACD, CCI and Momentum are all transformations of recent closing prices. When they agree, that is not five independent confirmations — it is one measurement expressed five ways, and the agreement is close to guaranteed. Meanwhile the screen has become unreadable and every entry now requires five conditions, so the ones that pass are the ones already well underway.
Real confluence combines genuinely different information:
Direction or trend — where the higher-timeframe bias sits.
Structure or location — whether price is somewhere a move can start from.
Momentum or participation — whether anything is actually behind the move.
Risk — where the idea is wrong, in price, before entry.Four inputs from four categories tell you more than fifteen oscillators. If you can't name which category each indicator on your chart belongs to, you have duplicates.
4. Survivorship bias in everything you were shown
You see the screenshots that worked. You do not see the same indicator over the flat month, the false-break week, or the market it was never tested on. Nobody posts those, and there's no conspiracy required — people post their best charts.
This is why a gallery of winning trades carries so little information, and why we don't publish one. A fair picture requires the losing sequences at the same size, on the same markets, with the same settings, and almost nobody shows that.
The only version you can trust is the one you generate yourself, on your market, over a window you chose before you looked.
What the ones that survive have in common
Finalised signals. A signal that can be withdrawn was never a signal.
Few, explainable inputs. You should be able to say what every setting does and why it exists.
Volatility awareness. Fixed-point stops and targets that ignore current range are wrong at least half the time.
Non-duplicated logic. Direction, location, momentum and risk — not five momentum readings.
Honest lateness. Confirmation costs you the first part of the move. Tools that claim otherwise are usually measuring the unclosed candle.
A defined invalidation. If nothing on the chart tells you where the idea is wrong, it isn't a trading tool, it's a decoration.The uncomfortable part
Even an indicator that passes all six will not rescue an account that risks too much per trade. Position sizing determines survival; the indicator determines only what you look at. That's why every tool page on this site links to the position size calculator and the risk of ruin calculator rather than to a results page.
If you want to judge any tool — ours included — run the verification method rather than reading another review.
We publish no win rates, no return figures and no backtest statistics — for our own tools or anyone else's — because no vendor can prove those numbers to you and you should not trade on a number you cannot verify. Everything above is educational. Trading carries a substantial risk of loss. SimpleAlgo is not affiliated with, endorsed by, or sponsored by TradingView. TradingView is a trademark of TradingView, Inc.
