Weekly brief · week of 23 September 2019

Week of 23 September 2019: Bitcoin and Ether lose about a fifth of their value

When Bitcoin and Ether fall about 20% in the same week, how concentrated was that fall in time, and how should a reader weigh the difference in their trading volume?

BTC fell from 10,028.05 to 8,043.82 (-19.79%) and ETH from 211.20 to 169.24 (-19.87%). Tuesday 24 September alone took 12.46% off BTC and 17.61% off ETH. BTC's weekly return ranked at the 1st percentile of the 84 weeks on record; ETH's at the 11th.

What the data shows. Both markets opened the week at their highs. BTC's high of 10,049.99 came in the first hour of Monday 23 September, 0.22% above its open of 10,028.05, and it closed at 8,043.82. ETH's high of 211.68 also came in that first hour, and it closed at 169.24. Most of the damage landed on Tuesday 24 September: BTC lost 12.46% that day and ETH 17.61%, with the week's largest hourly moves in the hour starting 18:00 UTC (-8.95% for BTC, -12.71% for ETH). ETH reached its low of 150.03 at 19:00 UTC on Tuesday; BTC's low of 7,750.00 came later, at 16:00 UTC on Thursday 26 September. Both closed in the lower third of their weekly range. The two moved together (hourly correlation 0.88) and the ETH/BTC ratio changed only -0.10%, so this was a shared move, not one asset breaking away. BTC's dollar volume of 3.1 billion was 0.92 times its prior twelve-week average; ETH's 0.88 billion was 1.87 times its own.

What it does not show. Price data cannot say what caused the Tuesday hour or who was on each side of it. The volume gap between the two assets has no explanation in this dataset, and nothing here indicates whether the low was final.

What a careful desk would check. Realised volatility, the size of the swings that actually happened, scaled to a yearly figure, was 95.5% for BTC and 127.3% for ETH. A position sized on the prior four weeks—three of which were down for BTC (-4.12%, 6.76%, -0.76%, -2.66%) and three of which were up for ETH (-8.47%, 6.13%, 4.33%, 11.72%)—would sit differently in each asset. Stop distances set before Tuesday would have been tested by hourly moves of 8.95% and 12.71%.

Funding. BTC perpetual funding was 0.01% per eight hours at its average, its maximum and its minimum alike, so the rate did not move during the week. That is a small, steady cost for long positions. No funding figures are available for ETH here.

What would change this reading

  • If BTC closes a later week back above its open of 10,028.05, the Tuesday hour reads as a shock that was absorbed.
  • If ETH volume stays near 1.87 times its prior twelve-week average while BTC stays below 1.0, attention has shifted between the two assets.
  • If the ETH/BTC ratio begins changing by more than -0.10% a week, the shared move has become a relative one.
  • If BTC funding moves away from 0.01% per eight hours, the steady cost that accompanied this week has ended.

Terms

realised volatility
how much the price actually swung over a period, expressed as a yearly percentage
percentile
the share of past weeks with a lower value; the 1st percentile means only 1% of the 84 weeks were worse
perpetual funding
a regular payment between long and short holders of a perpetual futures contract that keeps its price close to the spot price

Research notes on real market weeks, not investment advice. Sources: Binance spot BTCUSDT and ETHUSDT 1-hour candles and USD-M perpetual funding rates; weekly figures computed by WSTiger.

RISCON Labsrisconlabs.com