And what it costs to be out on the wrong day
Every scenario compounds the same daily closes. Sitting out a day means that day earns zero rather than being removed from the series, so the periods stay identical in length and only the participation differs. Perfect timing holds on every day that closed higher and sits in cash on the rest, which is the arithmetic ceiling rather than an achievable result.
Fees, spreads, slippage and tax are all excluded, which flatters the market-timing rows specifically: every one of them requires dozens or thousands of round trips that the buy-and-hold row never pays for. The real gap is wider than shown, in the direction that favours doing nothing.
The extreme days are ranked by daily close-to-close return over the full listed history, so the window depends on when the coin began trading rather than a fixed lookback. Days are UTC, matching the exchange candles.
Everyone who has held through a crash has run the counterfactual: sell at the top, buy back at the bottom, keep the difference. The arithmetic of that fantasy is easy to produce and comes out absurd — holding only on the days a coin rose returns a figure with more digits than there is money in the world. A result that large is not an ambitious target. It is the calculation telling you the assumption is impossible.
The version worth measuring is the realistic failure. Attempts to sell the top mostly end with being out of the market for a handful of days, and in an asset where returns are concentrated into a few sessions, that is expensive out of all proportion to the time involved. Missing twenty days out of several thousand is enough to turn nine years of Bitcoin into nothing.
The reason both cannot be had is timing, not skill. The largest single-day gains sit immediately beside the largest single-day losses, because violent rebounds are what follow capitulation. Bitcoin's best day in its listed history came the day after its worst. Anyone in cash for the crash was, in all likelihood, still in cash the following morning.
⚠️ Not investment advice. These are backward-looking counterfactuals computed on historical closes with no fees, spreads, slippage or tax, and past concentration of returns is not a promise about future ones. All decisions and risks are your own.
An amount that is not meaningful. Holding only on days that closed higher produces a figure with more digits than there is money in existence. The size of that number is the finding: it is the arithmetic ceiling, not a target anyone approaches.
Far more than the days suggest. Crypto returns are concentrated into a small number of sessions, so sitting out the twenty best days of several thousand is enough to erase most of the return — and for Bitcoin over its full listed history, enough to turn it negative.
Because the best days and the worst days are the same week. Violent rebounds follow capitulation, so the largest single-day gains sit immediately beside the largest single-day losses. Bitcoin’s best day in its listed history came the day after its worst.
For Bitcoin, you end up behind having simply held. That row is on the page because it is the realistic outcome of active timing — you do not get to keep the crash avoidance and the rebounds separately.
No, and that omission favours the timing scenarios. Every market-timing row requires dozens or thousands of round trips that buy and hold never pays for, so the real gap is wider than shown, in the direction of doing nothing.
No. A day sat out simply earns zero for that scenario. Every row covers exactly the same calendar period and the same number of days, so the comparison isolates participation rather than changing the window.
The window is each coin’s full daily history on the exchange, so it begins when the pair was listed. A newer coin therefore covers fewer days and fewer market cycles, which makes its extreme-day statistics less settled.