🏠 Home 🔍 Today's Scan 📰 Daily Report 📈 Hit Rate 📊 Base Rate ❓ Methodology 📚 Learn 🪙 Crypto Scanner 📋 All Tools ⏪ Investment Simulator 🧾 Tax Calculator🧮 Pension vs. Direct ⚡ Leverage ⚖️ Rebalancing 💹 DCA 📉 Averaging-Down B/E 💰 Dividend Calendar
← Back to the Learn hub
🎲 Statistical Traps · Verification

Multiple Testing
— Run 30 Indicators and One Wins by Chance

When you test a single indicator, "a 5% chance it's random" looks low. But test 30 of them at once and the story changes. This is why so many "we found a significant signal" claims are an optical illusion.

Written by Dawn · IT Engineer · Published
💡 Key takeaway — Try enough times and you'll get plenty of lucky hits too. A "significant finding" that doesn't disclose how many tests were run can't be trusted.

The Structure of the Problem

A common threshold in statistical testing is "less than a 5% chance this result happened by chance." For a single indicator, that's a reasonable bar.

But what happens if you test 20 completely useless indicators against that same bar? Each one has a 5% chance of passing purely by luck, so on average, 1 of them passes — even though none of the indicators actually carries any information at all.

Number Tested at OnceChance at Least One Useless Indicator Looks "Significant"
15%
1040%
2064%
3079%

Run 30 tests and 8 times out of 10, at least one fake finding sneaks in. And people tend to remember and publish only the ones that "hit."

Common Forms This Takes in Investment Decisions

  • Searching indicator combinations — tweaking RSI × moving average × volume conditions every which way and picking whichever combination performed best. With hundreds of combinations, one of them is bound to look good.
  • Parameter tuning — "returns go up if you use 28 instead of an RSI of 30." Continually adjusting a value is also repeating the test.
  • Trying different time windows — if it doesn't fit, adjust the window. This is also a way of increasing the number of attempts.

A rule found this way looks perfect in the past and falls apart in the future. That's because it was fitted to a random pattern in the data.

⚠️ Publication Bias — Failed attempts usually don't get published. There are plenty of posts saying "this indicator worked," but "I tried 29 and 28 failed" posts are rare. Without knowing the denominator, there's no way to tell if that one hit was skill or luck.

The Fix — Correction and Pre-Registration

① Multiple-Comparison Correction (BH/FDR)

This is a method that tightens the bar the more tests you run. DawnScan applies the Benjamini-Hochberg (BH) procedure to control "the share of fakes among everything published as significant." The ✓BH mark on the base-rate proof page is exactly this.

A signal that only passes ✓Wilson (its confidence-interval lower bound exceeds the base rate) and a signal that also passes the BH correction carry different levels of trust.

② Measure First, Apply Later

When DawnScan adds a new signal, it sets its weight to 0 and only logs it. Only once enough of a sample has accumulated and it passes testing does it get folded into the actual score. This breaks the cycle of "looking at the result and then rewriting the rule."

③ Requiring Consistency Across Multiple Labels

Looking at just one criterion makes it easy to pass by chance. That's why we require the same direction across all three of the peak, path, and final-return labels. It's much harder for chance to line up three times in a row.

📌 Summary — When you hear "we found a significant signal," ask "how many did you test?" If the answer isn't "one," a correction is needed. And real verification means checking the rule, once it's set, against newly arriving data.
📮 Daily US Market Morning Brief — We send an analysis of the previous day's top 10 US gainers (TOP10) and what they had in common, every day at 8am (KST). Telegram @dawnbrief · Free · No ads · Not stock recommendations.

Frequently Asked Questions

What is the multiple-testing problem?

It's the problem where, if you test several hypotheses at once, some get judged "significant" purely by chance even when none of them actually has any real effect. Set the bar at 5% and test 20 indicators, and on average 1 passes purely by luck. With 30, the chance that at least one fake finding sneaks in rises to roughly 79%.

What is the BH correction?

The Benjamini-Hochberg procedure is a method for controlling "the share of fakes (the FDR) among everything published as significant" when running multiple tests at once. It adjusts the bar based on how many tests were run, so a signal that passes it is more trustworthy than one that passed with no correction at all.

Does adjusting a parameter also count as multiple testing?

Yes. Trying an RSI threshold of 28 instead of 30 is also a new test. Adjusting a value repeatedly and picking whichever result looks best increases the number of chances something looks good purely by luck, in proportion to how many attempts you made. A rule found this way tends to fit the past well and fall apart in the future.

So how should verification be done?

① Count the number of tests and apply a correction ② set the rule first, then confirm it against newly arriving data ③ check whether it holds up consistently across multiple evaluation criteria. DawnScan logs a new signal at a weight of 0 only, and only folds it in once a sample has built up and it passes both all three labels and the BH correction.

Related Reading