How to Track a Crypto Sell-Off with Bounce Watch

COMPOUNDFREE RESEARCH · · UPDATED

During a crypto sell-off, open Bounce Watch to see whether Bitcoin and Ethereum are weakening together, how many smaller coins are falling, and how far the combined BTC/ETH gauge has dropped from its recent high.

How to read a Red Alert

Red Alert appears when all three market checks confirm weakness. The watch timer records when the first warning appeared and continues through temporary recoveries. Select a coin to compare its price with its frozen 24-hour low and see how far it fell below that reference in past declines. Those historical levels give you context while the market is moving.

See the dashboard and market checks in our 10 November Red Alert walkthrough.

Open Bounce Watch →

The research behind Bounce Watch

A falling cryptocurrency market can create opportunities, but buying too early can mean watching prices continue lower. CompoundFree’s latest research explores a specific question: can historical behaviour help us choose more informed prices at which to watch for a recovery?

We examined a year of market movements, measuring the duration of declines, the rebounds that followed and how far individual coins fell beneath their preceding 24-hour lows. Those findings form the basis of Bounce Watch, CompoundFree’s separate market research tool.

What we found

Between 4 November 2025 and 4 November 2026, our combined BTC/ETH price gauge recorded 218 completed swings of at least 2% down followed by at least 2% up.

Across those completed swings, the median decline from peak to low lasted approximately 11.1 hours. The median time from the low to a confirmed 2% recovery was approximately four hours.

These timings varied widely. They describe historical movements rather than a fixed market schedule.

We also examined 13 smaller coins after 74 separately sampled market sell-offs. Among 960 coin observations with sufficient follow-up:

Rebound from the subsequently identified lowReached within 24 hours
At least 2%93.3%
At least 5%51.9%
At least 10%13.5%

The median maximum rebound was approximately 5.1% within 24 hours.

These figures show the scale of recoveries that occurred. They measure moves from lows identified afterwards, so they are not trading win rates. The categories also overlap: a coin recovering 10% counts in all three rows.

Looking beneath the 24-hour low

Our next investigation focused on the falling period itself.

When a coin broke its previous 24-hour low, how much further did it typically fall before reaching its eventual trough?

We used completed five-minute closing prices. At the first close beneath the lowest close of the preceding 24 hours, we froze that reference level. This prevented the comparison from moving downward every time price made another low.

PopulationCompleted declinesDeclines breaking the reference lowAverage extension below itMedian extension
BTC/ETH market gauge2181512.47%1.66%
Our 13 smaller coins combined8,9833,4162.39%1.79%

The averages and medians apply to declines that broke the reference. Many completed declines never broke their preceding 24-hour low.

For illustration, if a frozen reference were £100, an extension of 1.79% would place price at £98.21. Individual coins have their own historical distributions; the combined figure would not replace their individual measurements.

The spread of outcomes matters. For the BTC/ETH gauge, the middle half of extensions ranged from 1.04% to 3.39% below the reference, while the largest reached 10.35% below it. An average therefore provides context, not a precise bottom.

The smaller-coin observations are correlated, and these measurements concern completed declines followed by recovery. Unfinished falls remain outside those completed-swing averages.

How the original research proposed testing Bounce Watch

Bounce Watch would bring market conditions and individual coin history together in one view.

The proposed market indicator would turn red when all three conditions were present:

Red would identify measurable weakness. It would begin a watch period rather than automatically trigger a purchase.

The page would provide a watchlist of up to ten eligible coins. The selection process would consider historical recovery behaviour, liquidity and results after realistic costs. A final ranking has not yet been established.

Selecting a coin would display:

InformationWhat it tells the user
Current priceWhere the coin is trading now
Frozen 24-hour referenceThe level recorded at its first break
Current distance below that levelHow far the present decline has extended
Historical average and medianHow comparable completed declines behaved
Sample count and downside rangeHow much evidence sits behind the figures
Recovery statusWhether price is still falling, bouncing or retesting

The frozen reference would remain visible throughout that episode. Missing or stale data would suspend the setup.

Why missing some bounces may be acceptable

The entry idea is to use each coin’s historical median extension to define a simulated limit-buy price beneath its frozen reference.

Some recoveries would begin before reaching that price. Those orders would remain unfilled, and those trades would be missed.

That is an intentional trade-off: waiting for a lower entry rather than chasing every rebound. Whether the trade-off improves results depends on what happens to the orders that do fill, including those filled during deeper, continuing declines.

Bounce Watch would also follow the recovery sequence: an initial bounce, a pullback and any subsequent higher low. This would allow a comparison between buying at a predefined lower price and waiting for measurable reversal confirmation.

Keeping the approach measurable

Historical measurements used for an entry must come only from information available before that entry. The same applies to choosing the watchlist.

Every eligible setup would be recorded: filled orders, unfilled orders, missed rebounds, failed recoveries and exits. Fees, spread, slippage, order expiry and position limits would be included in the paper-trading rules.

Earlier CompoundFree rebound and support-entry tests lost money after modeled costs. Bounce Watch is a new research proposal, with live trading disabled.

Its purpose is straightforward: show when the market is weakening, place each coin’s decline in historical context and make entry decisions measurable. Capital protection comes first, and an unfilled order can be a valid outcome.