Quantifying Address Reuse
Address reuse rocketed up from 25% to 35% over the course of this 2020-21 bull run to the highest level since 2013
Recently I’ve seen the Glassnode metric Accumulation Addresses floating around.
Here DilutionProof is careful to caveat that this metric relies on reused Bitcoin addresses. I saw this and was fairly skeptical flows from reused addresses would provide enough signal to generate a reliable onchain metric.
Surely the number of Bitcoin transactors reusing addresses is tiny, given preventing address reuse is probably the #1 privacy win and most wallets actively try to discourage it?
I’ve never actually seen someone run the numbers on this, so I decided to calculate it myself and was surprised. Currently 35% of unique Bitcoin addresses per day are used previously. See the notes1 for some detail on how I calculated this.
This is much higher than I thought. It makes me more confident in metrics like Accumulation Addresses but less confident in the median Bitcoiner out there sending UTXOs into previously used addresses. It’s a bit of a reality check that I am in a bubble of people who obsess over coin control best practices, while many people just let it rip with the same address again and again.
Address reuse spiked up to 50% (!) in late 2013 and then declined into a range between 25-30% depending on market conditions. Bull markets attract new users who are prone to address reuse which then declines again in bear markets.
Address reuse rocketed up from 25% to 35% over the course of this 2020-21 bull run to hit it’s highest level since that 2013 spike. The uptrend in address reuse during the 2017-18 bull market was comparatively small.
I hoped address reuse would be declining over time due to better wallet UI but that is not the case. If you’re out there reusing addresses, check out this guide to learn how to improve your Bitcoin privacy.
The Address Reuse data shown above is calculated as follows:
Numerator: Distinct receiving addresses per day that ever received an output in a prior block
Denominator: Distinct receiving addresses per day
Both numerator and denominator are filtered to exclude addresses that have received outputs in more than 100 transactions. The idea here is to ignore exchanges and other businesses that reuse addresses and focus on something closer to normal users.
14 day moving average applied to the time series to smooth out trends
Did you look at the distribution of the amounts sent to re-used addresses? Would be curious to see if this is different from the overall distribution.