NOTE: Methodology changed! Previously, we were using the 24 hour change % from prior day and summing up the change % by day of week in each month. If Bitcoin was going up at beginning of the month then going down at end of the month, this led to overall change % for the month and day near zero. Which could get misleading if just one week a change on Wed recorded +4% and the other three weeks recorded -1.25%. I noticed several near 0% for the last three or so months and when I looked at a few it was because on the whole, the gains and losses cancelled out due to bing in a side-ways market lately with short bursts in one direction or another to new range-bound trading zones.
So I made two changes:
- Every day’s change % is calculated as it’s HLC3 ((high + low + close) / 3) relative to the SMA of the HLC3 the preceding 7 days.
- The median of each day of week of each month is used rather than average. This picks the value near the middle of the collected values for that day of week in the month tallied and helps offsets the effect of an unbalanced swing or even one too balanced and cancelling itself out!
Here are the last 12 months through end of Feb 2019:
NOTE: The data is from BInance and the daily candle closes at midnight UTC and opens with first trade or 15 seconds past midnight (whichever occurs first).