Week of 16 December 2019: Bitcoin gains 5.36% while Ether loses 7.28%, a 12.64-point split
How much of the gap between Bitcoin's weekly rise and Ether's fall comes from when each coin moved, and what can prices and funding rates (payments between long and short holders of no-expiry perpetual futures) show about it?
The answer matters to a desk that holds both coins and sizes them as one risk. BTC rose from 7,119.60 to 7,501.44 (+5.36%), closing 98.4% of the way up its low-to-high range. ETH fell from 142.46 to 132.09 (-7.28%); its high came in the week's first hour. Both bottomed at 13:00 UTC on Wednesday 18 December. The ETH/BTC ratio (Ether's price in Bitcoin) fell 12.0%.
What the data shows. The two assets bottomed together. BTC's low of 6,435.00 and ETH's low of 116.26 both came at 13:00 UTC on Wednesday 18 December. From there they separated. BTC's largest hourly move (+4.18%) came at 22:00 UTC that day, inside its best day of +9.87%. BTC kept rising to a weekly high of 7,518.54 in the week's final hour. ETH's best day was the same Wednesday (+8.94%), but its high of 142.72 had come in the week's first hour, and it closed at 59.8% of its range. ETH's weekly range (the gap between high and low) was 22.76%, against 16.84% for BTC. Its realised volatility, the annualised size of its hourly swings, was 76.8% against 59.9%. Hourly correlation was 0.79, which the data labels close. Volume was light for both: BTC traded at 0.87 times and ETH at 0.79 times their prior twelve-week averages.
What it does not show. The data records the divergence but not its cause. Neither the candles nor the funding rates show who was buying or selling or how large positions were. ETH's weaker week therefore cannot be attributed to any specific flow. BTC remains 46.27% below its 52-week high and ETH 63.09% below. The weekly split does not show which asset will do better from here, and one week cannot forecast the next.
What a careful desk would check. BTC's 5.36% week ranks at the 77th percentile of the 96 weeks on record, while ETH's -7.28% sits at the 26th. In the prior four weeks BTC returned -18.81%, 7.11%, 1.6% and -5.21%, and ETH returned -23.86%, 7.67%, -0.19% and -5.3%. That gives BTC two down weeks and ETH three. A desk would check in later weeks whether the Wednesday joint low holds and whether the ratio steadies after its 12.0% drop.
Funding. BTC funding averaged -0.001% per eight hours, with a low of -0.0223%, so on average shorts paid longs a very small amount. ETH funding was positive throughout, averaging +0.0093% with a low of +0.0034%, so longs paid shorts. Each rate ran against its own price direction: shorts paid while BTC rose, and longs paid while ETH fell. The rates show which side paid the other, not how many positions were open or how large they were.
What would change this reading
- If BTC stays above its weekly open of 7,119.60 while ETH stays below its weekly high of 142.72, the split will have lasted beyond one week. Prices alone would still not show why.
- If ETH trades back down to 116.26 in a later week, Wednesday's joint low would not have held as a lower bound for ETH.
- If BTC funding stays near -0.001% per eight hours while price holds its gains, shorts would keep paying longs a very small amount. Funding alone does not show how many positions are open or how large they are.
- If dollar volume returns towards the prior twelve-week averages after this week's 0.87 and 0.79 ratios, later price moves would rest on more trading than this week's did.
Terms
- funding
- a payment exchanged every eight hours between long and short holders of a perpetual futures contract; when the rate is negative shorts pay longs, and when it is positive longs pay shorts
- perpetual contract
- a futures contract with no expiry date, kept close to the spot price by funding payments
- close location
- where the week's close sat inside its high-low range, as a percentage; 98.4% means the week ended near its top
- range
- the gap between the week's high and low prices, shown as a percentage
- realised volatility
- how large the actual hourly price swings were during the week, scaled to a yearly rate
- correlation
- how closely two prices moved together hour by hour, on a scale where 1 means they moved in lockstep
- ETH/BTC ratio
- the price of one Ether expressed in Bitcoin, showing which asset gained relative to the other
- percentile
- the share of past weeks with a lower value; the 77th percentile means 77% of the 96 weeks on record were lower