Ethereum’s past outflow charts can change when more exchange wallets are identified
Coin Metrics rebuilt Ethereum’s historical daily and hourly Standard Flow Metrics using updated exchange-address information, so past values may differ from earlier downloads. This creates a risk of look-ahead bias in backtests: today’s revised history may include wallet attribution unavailable to traders at the time. Its Point-in-Time data avoids later rewrites but reflects a different attribution method and has limited coverage. Coin Metrics disclosed no revision sizes or ETH strategy results, so their effect remains unmeasured. Analysts should preserve data vintages and compare pre- and post-rebuild values, while accounting for when inputs became available. Any test should hold the trading rule fixed and separate changes in data, signals, trades and returns. Exchange outflows alone do not establish buying or profitable trading.
