Week of 27 February 2023: Bitcoin and Ether fall near 5%, both closing near their weekly lows
When a moderate weekly decline concentrates in a single hour and both assets close near their lows, what does that hour imply about how risk limits set during calmer weeks were sized?
BTC fell from 23,554.85 to 22,430.24 (-4.77%) and ETH from 1,641.37 to 1,564.58 (-4.68%). Both made their weekly low and their largest hourly move in the hour starting 01:00 UTC on Friday 3 March, and both closed in the lower third of their weekly ranges. Funding stayed near zero.
What the data shows. BTC opened at 23,554.85, rose to 24,000.00 at 08:00 UTC on Wednesday 1 March, and fell to 21,971.13 at 01:00 UTC on Friday 3 March. ETH traced the same path: high 1,677.86 on Thursday 2 March, low 1,544.39 in that same Friday hour. In the hour starting 01:00 UTC on Friday, BTC lost 5.42% and ETH 5.49%, the largest hourly move of the week for both; Friday was also the worst day for both (BTC -4.73%, ETH -4.76%). Both closed in the lower third of their weekly ranges (22.6% for BTC, 15.1% for ETH). Hourly correlation was 0.96 and the ETH/BTC ratio changed 0.1%, so this was a broad move, not one asset separating from the other. Dollar volume was 39.2 billion for BTC and 4.83 billion for ETH, or 1.07 and 1.11 times their prior twelve-week averages.
What it does not show. The candles record where prices went, not why. Nothing here identifies who sold or whether the Friday hour held forced liquidation or ordinary flow. The weekly returns sit at the 22nd percentile for BTC and 29th for ETH against 263 weeks of history, so these were below-average down weeks in the sample rather than extreme ones. Three of the prior four weeks were down; that is a pattern in price, not an explanation.
What a careful desk would check. Realised volatility (the size of the swings that actually happened, scaled to a yearly figure) reached 51.4% for BTC and 54.0% for ETH. BTC closed 53.24% below its 52-week high and ETH 55.86% below. A position sized for an ordinary week could find its largest single-hour loss concentrated in that Friday hour. A desk would compare stop distances and margin buffers with a 5.42% hourly move rather than with the average week, and ask whether exits could have been filled inside that hour at all.
Funding. BTC perpetual funding averaged 0.0044% per eight hours, ranging from -0.0042% to 0.01%; ETH averaged 0.0056%, from -0.0009% to 0.01%. Rates near zero mean the payments between long and short holders were small; they do not show how large the positions behind them were, so crowding cannot be measured here.
What would change this reading
- If a later week trades below 21,971.13 for BTC and 1,544.39 for ETH, the Friday low was a step down rather than a floor.
- If price recovers above the weekly opens of 23,554.85 and 1,641.37 on volume near 1.0 times the prior twelve-week average, the week reads as a dip inside a range rather than a change of trend.
- If funding turns clearly positive while price keeps falling, longs are paying to hold positions that are no longer rising, and that carry becomes measurable against the week's 5.42% hourly move.
- If the ETH/BTC ratio moves well beyond its 0.1% change while hourly correlation stops reading as very close, the broad move has become a relative one, which changes which hedge matters.
Terms
- realised volatility
- how much the price actually swung over a period, expressed as a yearly percentage
- hourly correlation
- how closely two prices moved together hour by hour, from -1 to 1
- funding rate
- the periodic payment between long and short holders of a perpetual futures contract; positive means longs pay shorts