This article is educational. It is not personalised investment advice, not a claim that any signal service is profitable for any individual, and not a recommendation to buy or sell any security.
“Do stock signals work?” The honest answer is: some do, most don’t, and the difference is rarely what the marketing points at. A signal “works” only if it clears three bars at once: positive expectancy net of costs, a result that reproduces out of sample, and a human disciplined enough to actually follow it. Plenty of services clear the first in a curated backtest and quietly fail the other two. Here is how to tell the difference.
“Work” means positive expectancy, not a good week
A signal that works is not one that was right last week, or that produced a screenshot of a 300% winner. It is one with positive expectancy: the average outcome per signal, across the whole population, after fees and slippage. A service can always show you its winners and bury a longer tail of losers. The only number that survives that game is the average over every signal, winners and losers together.
A worked example makes the point. Suppose a system takes 100 trades and wins only 40 of them. If the average winner makes +20% and the average loser costs 7%, the winners add 800 points and the losers subtract 420: a net of +380 across 100 trades, or roughly +3.8% per trade before costs. Now “improve” the hit rate to 65% but cut the winners to +3% while letting losers run to 12%: the same arithmetic turns negative. Accuracy went up; the edge went away. This is why a 90% “win rate” in an advert is a warning sign, not a selling point. The math is laid out in profit factor explained and why hit rate misleads.
The three tests a signal has to pass
Strip away the marketing and the bar is concrete. A signal worth following has to clear all three of these, not just the first.
- Positive expectancy after real costs. Commissions, the bid/ask spread, and slippage between the signalled price and the fill you actually get all subtract from the edge. A thin backtest edge often vanishes once those are charged honestly. The right question is not “is it profitable on paper” but “is it profitable after the costs a real account pays.”
- Reproducible out of sample. The backtest has to land on the same number when re-run, survive the loss of its few best trades, and be built on a survivorship-free universe. It also has to be discounted for how many variations were tried before this one looked good: test 200 rules and one will shine by luck alone. The traps live in why backtests lie, overfitting, and survivorship bias.
- Followed with discipline. Even a genuine edge returns nothing to the person who overrides it, skips the losing stretch, or sizes erratically. The edge lives in the average over many trades, so taking only the comfortable ones is a different (usually worse) strategy.
Why most signal services fail
Almost every failure is one of two kinds, and they map onto the first two tests above.
- No real edge. The track record looks brilliant because the history was curved-fit to its own backtest, tuned on the names that happened to survive, or never checked on data the rule had not seen. A headline performance figure has to be deflated for that selection (Bailey & López de Prado, 2014). The base rate is unkind: most professional active managers fail to beat a plain index over a decade (S&P SPIVA), which is the same “edge that did not survive contact with out-of-sample” story at industry scale.
- The human does not follow it. The most active individual investors underperform the most (Barber & Odean, 2000), and the average fund investor earns meaningfully less than the funds they hold because they buy and sell at the wrong time (Morningstar’s “Mind the Gap” studies). That gap is a discipline problem, not a signal problem, and no signal fixes it for you.
What it looks like when a signal does work
A working systematic signal is usually less accurate than you would guess and more profitable than its hit rate suggests, because the payoff is asymmetric: small, capped losses and a few large winners. Shishin’s own published backtest is a case in point. It wins under half of its trades, roughly four in ten, yet compounds, precisely because the winners are allowed to run several times the size of the losers. The full five-year record is survivorship-free, reproducible to the dollar, and shows every trade, winners and losers together, not a highlight reel: the figures live on the track record. The same locked configuration now runs live in public on a paper account, so the out-of-sample test is happening in the open rather than being asserted, which is the point of running what was tested.
How to vet any signal service in five questions
- Is the edge positive after commissions, spread, and slippage, not just on frictionless paper?
- Is the record survivorship-free and reproducible, and does it survive losing its best few trades (see leave out the winners)?
- Does it show the losers and the full distribution, or only curated winners? The detailed checklist is in how to vet a track record.
- Is there an out-of-sample or live record, or only an in-sample backtest?
- Is the cadence and drawdown shape one you can actually stick to through a losing stretch?
Where Shishin stands
We are a research publisher, not an adviser, so “does it work” is a question we try to let you answer rather than assert. The daily signals are proof of life; the real evidence is the multi-year backtest plus the live record, shown in full with the losers included. What you do with the research is your decision. A signal only ever improves the odds. It never removes the risk.
Sources & further reading
- Barber, B. M. & Odean, T. (2000). “Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors.” Journal of Finance, 55(2), 773 to 806.
- Jegadeesh, N. & Titman, S. (1993). “Returns to Buying Winners and Selling Losers.” Journal of Finance, 48(1), 65 to 91: a profitable edge with a sub-50% hit rate.
- Bailey, D. H. & López de Prado, M. (2014). “The Deflated Sharpe Ratio.” Journal of Portfolio Management, 40(5), 94 to 107: why a headline figure must be discounted for how many variants were tried.
- S&P Dow Jones Indices, SPIVA U.S. Scorecard: the share of active managers underperforming their benchmark over long horizons.
- Morningstar, Mind the Gap (annual): the gap between fund returns and the returns investors actually realise, a measure of the discipline problem.