Week of 3 September 2018: Bitcoin loses 14% and Ether 34% as both close near the week low
When Bitcoin and Ether fall in the same week but Ether loses more than twice as much, what does the gap show, and what can a week of price data not explain?
BTC fell from 7,302.00 to 6,250.81 (-14.40%) and ETH from 295.51 to 195.81 (-33.74%). Both closed in the lower third of their weekly ranges after the low came later than the high. ETH's weekly return ranked at the 0th percentile of 29 recorded weeks, BTC's at the 10th.
What the data shows. BTC opened at 7,302.00, reached 7,410.00 on Tuesday 4 September at 12:00 UTC, and fell to 6,111.00 on Saturday 8 September at 20:00 UTC — a 21.26% range, with the low after the high, so the week trended down rather than reversing. Wednesday 5 September was the worst day for both (BTC -8.96%, ETH -19.86%) and held BTC's largest hourly move (-3.94% in the hour starting 09:00 UTC). ETH set its high, 296.62, in the week's first candle on Monday 00:00 UTC and its low, 185.02, on Sunday 9 September at 03:00 UTC. Both closed in the lower third of the weekly range (BTC at 10.8%, ETH at 9.7%). Dollar volume ran above the prior twelve-week average for both (1.12 times for BTC, 1.28 times for ETH).
What it does not show. Higher volume records more trading, not the reason for it. Nothing here identifies who sold or whether the selling was forced. The largest hourly move of the week for ETH was a gain (+6.46%, Sunday 09:00 UTC), which shows a bounce occurred but not whether it held.
What a careful desk would check. Percentiles rest on 29 weeks of history, a thin base: an extreme rank here means less than it would against years of data. Realised volatility, the size of the actual swings scaled to a yearly figure, was 65.9% for BTC and 125.3% for ETH. A position sized for the prior four weeks, in which BTC rose in three weeks, would be too large for a week like this.
Ether against Bitcoin. The two assets moved closely but not identically (hourly correlation 0.84), and the ETH/BTC ratio fell 22.6%, so this was a broad decline with a distinct ETH leg on top. The weeks before had diverged: BTC rose in three of four (+2.68%, +3.43%, +8.99%), ETH fell in three of four (-22.23%, -5.83%, -8.36%). At the close BTC sat 46.68% below its 52-week high, ETH 83.11% below.
What would change this reading
- If the ETH/BTC ratio stops falling in later weeks, the 22.6% drop reads as a one-week divergence rather than a lasting change in relative performance.
- If BTC trades back above this week's open of 7,302.00 while volume falls to its prior twelve-week average, the week reads as a shock that was absorbed.
- If a later week closes below this week's low of 6,111.00 for BTC or 185.02 for ETH, the Sunday bounce was a pause inside the decline.
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
- percentile
- the share of past weeks with a lower value; the 0th percentile means no recorded week was worse