Does Bitcoin bounce after a 20-day low?
No, not measurably. BTC cleared its cost 58.1 % of the time against 52.1 % on an ordinary day. The gap is inside what chance produces at this sample size.
Measured on daily data from 2020-11-02 to 2026-10-12, after the cost of trading. Recomputed every time this page loads.
When BTC makes a 20-day low, does BTC outperform its baseline over the next 5 days?
A question only means something if it is stated precisely enough to be wrong. This is the form the engine measured.
Buying the dip assumes a fall is an opportunity rather than information. A 20-day low is the simplest way to define one.
86 signal days since 2020-11-02. On 50 of them BTC rose more than break-even over the next 5 days (58.1%). On an ordinary day the same LONG works 52.1% of the time — so the condition adds +6.0 points.
Typical move on a signal day: +0.73% against +0.24% on an ordinary day. Worst signal day: -28.13%.
Chance: a coin that lands LONG-right 52.1% of the time, flipped 86 times, gives 50 or more hits in about 28% of runs (p = 0.2817). That is not distinguishable from luck.
Consistency: first half of the period 67.4%, second half 48.8% — not both above baseline; one half carries it.
What it would take: at this hit rate about 267 signal days for p < 0.05 and about 1101 to survive the family of 702 tests. There have been 86.
A relationship that only appears in one year is usually that year, not a relationship.
the hit rate over the next 30 signal days falls to the baseline (52.1 %) or below — the relationship disappears out of sample
Stated before the next observation, not after. The page recomputes, so if this stops being true the answer here changes with it.
Change the condition, the asset, the direction or the holding period and measure it against the same six years.
Open this in the tester