Statistics · Cycles & signals

Bitcoin RSI Bottom Signal

The weekly RSI(14) reflects price action only, with no fundamentals. Since 2011 it has exited its normal band downward 7 times (below 35). In 60% of the resolved cases the cycle bottom fell within ±13 weeks (median 3 weeks), but not every signal was a bottom. RSI currently reads 59.8 (in the normal band). With n=7 the sample is very small; this describes the past and is not a forecast.

Weekly RSI now
59.8
normal band
Signals since 2011
7
RSI below 35, one square per signal
2011
2014
2018
2020
2022
2026
2026
Median 6-month return
+38 %
3 of 6 higher afterwards
Bottom ±13 weeks
3 of 5
60% · median 3 weeks

Squares: filled = higher afterwards or bottom within ±13 weeks, empty = lower or further away, dashed = still open.

Weekly RSI across history

The weekly RSI almost always oscillates inside the "normal band" (35-70, grey). Since 2011 it has dropped below only 7 times (orange dots); in 60% of the resolved cases the cycle bottom fell within ±13 weeks. So the chart shows how rare the signal is and when it fired; what price did next is in the table below.

70503530201020122014201620182020202220242026OVERBOUGHTNORMAL BANDOVERSOLDNOW · RSI 60

Hover the chart for date, RSI & price.

Bold line = weekly RSI(14) · faint line = price (log) · tinted zone below 35 = oversold · ● = signal.

What happened next (base rates)

From 7 signals since 2011: the share that was higher after X months, with a Wilson 95% interval.

after 1 month
4 of 7 higher
median +0.4 % · range -32 % to +45 %
0%50% chance100%
Wilson 25-84%
after 3 months
4 of 7 higher
median +2.0 % · range -29 % to +99 %
0%50% chance100%
Wilson 25-84%
after 6 months
3 of 6 higher
median +38 % · range -46 % to +128 %
0%50% chance100%
Wilson 19-81%
after 12 months
3 of 5 higher
median +75 % · range -38 % to +1,007 %
0%50% chance100%
Wilson 23-88%

Squares: one per signal, filled = higher afterwards, empty = lower, dashed = not long enough ago. Bar: hit rate (tick) with Wilson 95% range; the line at 50% is a coin flip. Return from the signal's close; "median" = middle value, not average.

The individual signals

WeekPriceto bottom1·3·6·12M
November 14, 2011$2.20+0 %
1M higher, 3M higher, 6M higher, 12M higher
September 22, 2014$375-44 %
1M lower, 3M lower, 6M lower, 12M lower
November 19, 2018$3,934-19 %
1M higher, 3M lower, 6M higher, 12M higher
March 9, 2020$5,349+0 %
1M higher, 3M higher, 6M higher, 12M higher
May 16, 2022$30,320-46 %
1M lower, 3M lower, 6M lower, 12M lower
January 26, 2026$77,057-23 %
1M lower, 3M higher, 6M lower, 12M open
June 1, 2026$63,181-6 %
1M higher, 3M higher, 6M open, 12M open

* bottom episode not yet resolved (current cycle). Squares: price after 1, 3, 6 and 12 months, filled = higher, empty = lower, dashed = still open.

Methodology & data

Weekly closes from our own price database (last price of each ISO week, USD as the historically traded price, since 2010). RSI(14) using Wilder smoothing. Signal = week the RSI falls from ≥35 to <35; one signal per episode (reset only at RSI 50), detected from 2011. Return base rates use the nearest weekly close to the +1/3/6/12-month date; the hit rate is shown as a fraction with n, plus the Wilson 95% interval, which is more honest on small samples than a normal approximation. "Bottom" = lowest weekly close until recovery to the pre-signal high. Euro values at the ECB daily rate. With n=7 the sample is very small; the figures are illustrative.

Frequently asked questions

01What is the RSI(14)?

The Relative Strength Index compares the price gains of the last 14 periods with the losses, here on weekly closes. Gains and losses are averaged separately (Wilder smoothing) and put in relation; the result lies between 0 and 100. 50 means both balance out. Above 70 the price has risen sharply in a short time, very low values show heavy selling. The RSI measures price movement only, not its cause.

02What is the RSI bottom signal?

The weekly RSI(14) rarely leaves its "normal band" (35-70) to the downside. We mark every week it falls below 35 (one signal per episode, reset at RSI 50) and show what price did afterwards.

03Why is the threshold 35?

35 is a threshold we chose: at 30 the 2011 bottom and the 2020 Covid crash would be missing, at 40 the signal fires in almost every correction.

04How reliable is this?

There have been only 7 such signals since 2011, a very small sample. The base rates are illustrative, not statistically robust, and several signals can share the same cycle bottom. Past patterns need not repeat.