What Are Trading Signals? How to Tell If One Works

We tested 20 simple rules on SPY (the S&P 500 exchange-traded fund, or ETF). Each one held SPY only while its price was above an average of its recent closes. The best rule on 1993 to 2009 prices made 8.52% a year there. Holding SPY made 7.49%. On 2010 to 2026 prices, the same rule made 10.12%, while holding made 14.30%.

That rule is a trading signal: a rule that turns data into a buy, sell or hold. This guide explains the main kinds, then tests one on data it had never seen. It ends with a checklist for judging any signal or signal service.

A trading signal is a rule that turns data into an instruction: buy, sell or hold. The data can be prices, trading volume or economic figures. The rule is written so that two people would reach the same answer on the same day.

A signal is one part of a system. What Is Rules-Based Investing? adds the other parts, such as how much to buy and when to sell.

A backtest runs a rule over past prices as if it had been traded. What Is Backtesting? explains the method and its traps.

The trap that matters most here is choosing the best rule from many. The best of many looks good partly by luck. The fix is to split the data:

The rule holds SPY after any day it closes above its moving average. Otherwise it holds 3-month Treasury bills, which are US government loans. Twenty lengths were tested, from 10 to 200 trading days in steps of 10.

The signal uses each day's close. The position starts at that close and runs to the next one. Dividends are counted, while costs are left out unless stated.

Chart 1 shows each length's yearly return from 11 November 1993 to 31 December 2009. The 200-day average came first, at 8.52% a year. Holding SPY made 7.49%. Only 2 of the 20 lengths beat holding: the 200-day and the 170-day.

The 200-day rule also fell less. Its worst fall from a peak was 22.1%, against 55.2% for holding SPY.

Chart 2 runs the same 20 rules from 31 December 2009 to 6 October 2026. The 200-day rule kept first place, at 10.12% a year. But holding SPY made 14.30%. None of the 20 lengths beat it.

So the ranking held, while the lead over holding did not. $100 in the rule grew to $503. In SPY, it grew to $940. With a cost of 0.1% on each switch, the rule made 9.53% a year.

The rule still fell less. Its worst fall was 20.5%, against 33.7% for holding.

A base rate is what happens on all days, with no signal. A useful signal picks days that beat it.

Chart 3 shows SPY's average daily return, scaled to a year by multiplying by 252 trading days. A simple average of daily returns runs above a compound yearly return. Out of sample, for example, holding SPY made 14.30% a year compounded, while its average day, scaled up, gives 14.9%.

In sample, SPY rose at 11.7% a year on days the 200-day signal said hold. On days it said cash, the figure was 3.6%. The base rate was 9.3%.

Out of sample, the gap reversed. Hold days averaged 12.0%, while cash days averaged 30.0%, against a base rate of 14.9%.

The cash days were few, 663 of 4,215, but wild. All 10 of SPY's best days from 2010 to 2026 fell on them. So did 9 of its 10 worst days. The signal still dodged deep falls, but it no longer picked better days than the base rate.

A signal is a written rule, so it can be checked against the list above. For the 200-day signal, the checks gave a mixed answer. It cut SPY's worst fall in both periods. After 2010, it earned less than holding.

What Is RSI? and What Is MACD? run the same checks on those two signals.

A trading signal turns data into buy, sell or hold. Judge it on data it has not seen, after costs, against simply holding. In our test, the best moving-average rule kept its rank after 2010, but lost its lead over holding.

This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.

The same approach runs through everything else we publish:

Good places to start on the website:

Not by default. Each one has to be checked on data it was not built on. The best of 20 moving-average rules on SPY beat holding from 1993 to 2009. From 2010 to 2026, it made 10.12% a year, against 14.30% for holding SPY. Its worst fall stayed smaller, at 20.5% against 33.7%.

A trading signal is worth it only if it passes a few checks. Look for a live record, results on unseen data, costs counted, few rules tried, enough events and a stated worst fall. Compare every result with simply holding. In our SPY test from 2010 to 2026, costs of 0.1% a switch cut the 200-day rule from 10.12% to 9.53% a year.

Holding a fund only while its price is above its 200-day moving average is a simple trend signal. From 1993 to 2009, it made 8.52% a year on SPY, against 7.49% for holding. Other examples include the golden cross, the Relative Strength Index (RSI) and the Moving Average Convergence Divergence (MACD) indicator.

Quantitative signals are trading signals built from numbers with a fixed formula, so a computer can apply them. Prices, trading volume, company accounts and economic data can all feed one. Each signal turns its inputs into buy, sell or hold. Like any signal, it should be tested on data it was not built on.

DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.