Do popular trading sayings hold up? We tested 34 of them
Golden cross, sell in May, buy the dip, cut your losses: each saying is turned into a precise rule, run on up to 16 years of daily data for ten markets, and given one of four verdicts. Data through 2026-11-10.
The scorecard
- Holds 1 The effect shows up in most markets, passes the significance guard, and has the same sign in both halves of the history.
- Partly 11 It holds in some markets or periods, leans the right way without proof, or helps on risk but not on return.
- Does not hold 19 No effect we can tell from chance, or the opposite effect.
- Not enough data 3 Too few independent events or observations to judge either way.
Jump to a topic
Moving averages
Calendar and seasonality
“Sell in May and go away: stocks do better from November to April than May to October.”
+1.0%/month vs +1.4% for the rest (t = -0.7).
“Stocks rally in the last five trading days of December and the first two of January.”
+0.05%/day vs +0.06% for the rest (t = 0.0).
“January is the best month for stocks.”
+1.4%/month vs +1.2% for the rest (t = 0.2).
“The best days for stocks are around the turn of the month.”
+0.07%/day vs +0.05% for the rest (t = 0.5).
“Monday is the worst day of the week for stocks.”
+0.06%/day vs +0.06% for the rest (t = 0.2).
“September is the worst month of the year for stocks.”
-0.5%/month vs +1.4% for the rest (t = -1.7).
Indicators and oscillators
“Buy when RSI drops below 30 (oversold), sell when it rises above 70 (overbought).”
10-day return after the signal +0.1% vs +0.2% on all days; clear in 0 of 10 markets.
“Buy when MACD crosses above its signal line, sell when it crosses below.”
Better Sharpe in 6 of 10 markets (median 0.37 vs 0.29).
“Price touching the lower Bollinger Band is a buy; upper band is a sell.”
10-day return after the signal -0.1% vs +0.2% on all days; clear in 0 of 10 markets.
“After three down days in a row, the market bounces.”
3-day return after the signal 0.0% vs +0.1% on all days; clear in 0 of 10 markets.
Dips and volatility
“Buy the dip: after a down day or pullback, prices bounce.”
21-day return after the signal +1.0% vs +0.7% on all days; clear in 1 of 9 markets.
“When the VIX is high, it is time to buy (be greedy when others are fearful).”
63-day return after the signal +5.7% vs +1.2% on all days; clear in 1 of 9 markets.
“Markets take the stairs up and the elevator down.”
Negative skew in 5 of 6 markets (-0.43 median).
Trend and momentum
“The trend is your friend: buy what is going up, sell what is going down.”
Better Sharpe in 6 of 10 markets (median 0.34 vs 0.22).
“Buying at new highs works: new highs lead to more new highs.”
63-day return after the signal +3.6% vs +2.2% on all days; clear in 0 of 9 markets.
“Buy last year's winners and sell last year's losers (12-1 momentum).”
Beat the benchmark on Sharpe in both windows in 0 of 2 cases.
“Don't catch a falling knife: never buy something in a downtrend.”
21-day return after the signal -0.7% vs +0.7% on all days; clear in 0 of 6 markets.
“Buy 20-day or 55-day breakouts, the turtle way.”
Better Sharpe in 5 of 10 markets (median 0.29 vs 0.30).
Risk management
Gold and commodities
“Gold is a safe haven: it goes up when stocks crash.”
On 137 event days: +0.01% vs +0.03% on all days (t = -0.2).
“Oil and stocks move together: oil falling means stocks fall.”
Correlation 0.30 (0.43 / 0.22 by half).
“Natural gas rallies into winter: buy in autumn, sell in winter.”
-3.7%/month vs -0.4% for the rest (t = -1.6).
“The gold/silver ratio reverts: when it is high, silver is the buy.”
63-day return -1.4% vs +0.7% on all days (t = -1.4).
“When the dollar falls, gold rises.”
Correlation 0.02 (0.02 / 0.01 by half).
Crypto
“Bitcoin follows a four-year cycle around the halving: prices peak 12-18 months after each halving.”
At most 3 independent events: too few to judge.
“Bitcoin is digital gold: it holds up when stocks fall.”
On 99 event days: -3.07% vs +0.25% on all days (t = -4.6).
“Bitcoin pumps in October and Q4 ('Uptober').”
At most 13 observations: too few to judge.
“Ethereum follows Bitcoin: when BTC moves, alts follow a few days later.”
Best lagged correlation 0.049 at 2 day(s).
Portfolio and investing habits
“Missing the best 10 days wrecks your returns, so time in the market beats timing the market.”
Missing the best 10 days costs a median 36% of total growth.
“Dollar-cost averaging beats investing a lump sum.”
DCA ahead on a median 50% of start dates.
“In a crisis all correlations go to 1, so diversification fails when you need it.”
Pairwise correlation 0.25 in stress vs 0.16 otherwise.
How we judge
Every saying is written as an exact rule or condition. Rules are long-or-cash on daily closes: a position decided at the close of day t earns the return of day t+1, indicators use only data up to day t, results are before trading costs, and cash earns nothing. Event studies compare returns after a condition with the average for the same market, using non-overlapping events.
A result counts as statistically visible only with |t| of at least 2 and at least 20 independent events; anything weaker is labelled likely noise. Every test is split into two halves of the market's history and a claim can only hold if both halves agree. A claim holds when at least 60% of the usable markets show the effect; it is partly true when only some do, when the direction is right in at least 70% of five or more markets without proof, or when it helps on risk but not on return. With fewer than 10 events a market is not usable, and when fewer than half of the markets are usable the verdict is not enough data.
Prices are daily closes of ten markets (S&P 500, gold, silver, oil, natural gas, commodities, agriculture, EUR/USD, Bitcoin, Ether) from 2010 (Bitcoin 2014, Ether 2017). Several claims are about stocks, and we have one stock market, SPY, so they are decided on SPY alone and the other markets are shown for context. The page sources of each saying are as commonly cited and were not independently verified.
Descriptive statistics from historical data, before costs. Not a forecast and not investment advice.