Research · test #10
Is there a best month for altcoins? No edge
"Altcoins pump in November and dump in June" is repeated every year. We took six of the biggest altcoins, wrote the rules down first and shuffled their monthly returns 100,000 times to see whether any calendar month stands out more than chance allows. None does. One curiosity - June - is recorded as a hypothesis and will be checked on new data in 2027.
In short: Ethereum, BNB, XRP, Solana, Dogecoin and Cardano, 73-109 completed months each. The best month of each coin averaged between +18% and +102%, driven by a few huge months (medians are far lower); shuffled data produces a best month that good in 20% to 71% of cases. No coin passed. June was negative in 43 of 49 coin-months.
The rules (fixed before the run)
- Data: Binance spot monthly candles for ETH, BNB, XRP, SOL, DOGE and ADA against USDT; return = close of the last day divided by open of the first day, minus one; completed months only, up to September 2026, without each coin's listing month.
- Test, per coin: average return of each of the 12 calendar months; the statistic is the best of the twelve. 100,000 times we shuffled the coin's monthly returns between calendar months (keeping the number of observations per month) and counted how often the best average was at least as high as the real one. That share is the p-value corrected for checking twelve months.
- A coin passes only if that p-value is below 0.83% (5% divided by six coins), the best month is positive in both halves of the coin's history, and it has at least six observations.
- Bitcoin is covered by the Uptober test.
Result: no coin passes
| Coin | Months | Best month | Average | Median | Up years | p (corrected) | Both halves positive |
|---|---|---|---|---|---|---|---|
| ETH | 109 | Apr | +18% | +7% | 6/9 | 0.71 | no |
| BNB | 106 | Feb | +49% | +5% | 5/9 | 0.56 | no |
| XRP | 100 | Nov | +48% | -11% | 3/8 | 0.20 | yes |
| SOL | 73 | Jan | +47% | +9% | 3/6 | 0.62 | yes |
| DOGE | 86 | Jan | +102% | +4% | 4/7 | 0.43 | no |
| ADA | 101 | Feb | +32% | -6% | 3/8 | 0.71 | no |
Every best-month average is big and every p-value is far above the bar (0.0083). Look at the medians: for XRP and Cardano the "best" month has a negative median - the average comes from one or two huge years. Each coin's full year-by-month table: Ethereum, BNB, XRP, Solana, Dogecoin, Cardano.
One curiosity: June
June closed higher in only 6 of 49 coin-months (12%), was the worst month for five of the six coins and was negative for most coins in almost every year:
| June | ETH | BNB | XRP | SOL | DOGE | ADA |
|---|---|---|---|---|---|---|
| 2026 | -22% | -23% | -22% | -11% | -28% | -39% |
| 2025 | -2% | 0% | +3% | -1% | -14% | -17% |
| 2024 | -9% | -2% | -8% | -11% | -22% | -12% |
| 2023 | +3% | -22% | -8% | -9% | -7% | -23% |
| 2022 | -45% | -32% | -21% | -26% | -23% | -27% |
| 2021 | -16% | -14% | -32% | +8% | -22% | -20% |
| 2020 | -3% | -10% | -13% | -9% | +12% | |
| 2019 | +9% | -2% | -10% | -8% | ||
| 2018 | -21% | +3% | -23% | -38% |
Why we do not call it a finding. It is really nine Junes, not 49: the six coins move with bitcoin, so one bad June hits all of them at once. The test of the worst month, with the same shuffling and correction as above, gave p between 0.05 and 0.86 across the coins - no coin passes. Bitcoin itself fell in six of the nine Junes. The pattern could be real, or one of the twelve months that always looks special in a short history.
What this test does not say
- Six coins with 6-10 observations per calendar month give a test of low power. "No month passes" does not prove there is no seasonality; it shows that this history cannot support it.
- The coins are not independent (they follow bitcoin), so six coins are less than six separate checks. The correction for six coins is therefore conservative.
- Only coins that survived to today and are large today are included, which inflates average returns (not necessarily the calendar pattern).
- Monthly candles are UTC calendar months; no trading rule or costs are involved - this is a statistical test of returns, not a strategy.
Data and method
Rules were committed before any data was loaded, and the code before the run; the June hypothesis was added afterwards as a dated addendum and kept out of the verdict. Data: Binance public API, monthly candles. Tables: open data. Method: methodology. Same approach as our other tests.
FAQ
Is there a best month for altcoins?
Not one that beats chance. For Ethereum, BNB, XRP, Solana, Dogecoin and Cardano the best calendar month averages between +18% and +102%, but with 6-10 observations per month, shuffled data produces a best month that good in 20% to 71% of cases. None of the six passed.
Is June really the worst month for altcoins?
In our data June was negative in 43 of 49 coin-months and the worst month for five of six coins. But that is nine Junes and the coins move together, and the test of the worst month did not pass. It is a hypothesis; we will check it on June 2027.
Why can altcoin seasonality not be proven?
Each calendar month appears only 6-10 times in these coins' history and a few huge months decide an average. With twelve months checked, one of them always looks special. A pattern needs more years or a prediction made in advance.
How was the test done?
Monthly returns were shuffled between calendar months 100,000 times per coin; we counted how often the best of twelve month averages was at least as high as the real one. A coin passes only if that share is below 0.83%, the best month is positive in both halves of the history and has at least six observations.
Is this financial advice?
No. It is a statistical test of past monthly returns. Past patterns do not predict future months.