If you're relying on historical exchange data to backtest your ETH trading strategies, pay close attention to this data update. 📉
Coin Metrics recently rebuilt its historical Ethereum Standard Flow Metrics from genesis. While identifying newly discovered exchange wallets helps reconstruct past supply movements with higher precision, it introduces a major pitfall for traders: revision bias.
Charts downloaded today reflect entity knowledge gathered long after those trades occurred. If you backtest using today's polished metrics, your model is essentially trading on "future" address data that wasn't available at the time. To avoid distorted signal testing, analysts need to isolate data vintages or rely on Point-in-Time (PIT) metrics.
For algorithmic traders and quantitative researchers, missing this distinction can lead to overly optimistic strategy performance that fails in live markets.
How do you handle look-ahead bias when testing on-chain metrics—do you rely on Point-in-Time data, or stick to simple price action?
#Ethereum #Crypto #Onchain #tradingStrategy
Coin Metrics recently rebuilt its historical Ethereum Standard Flow Metrics from genesis. While identifying newly discovered exchange wallets helps reconstruct past supply movements with higher precision, it introduces a major pitfall for traders: revision bias.
Charts downloaded today reflect entity knowledge gathered long after those trades occurred. If you backtest using today's polished metrics, your model is essentially trading on "future" address data that wasn't available at the time. To avoid distorted signal testing, analysts need to isolate data vintages or rely on Point-in-Time (PIT) metrics.
For algorithmic traders and quantitative researchers, missing this distinction can lead to overly optimistic strategy performance that fails in live markets.
How do you handle look-ahead bias when testing on-chain metrics—do you rely on Point-in-Time data, or stick to simple price action?
#Ethereum #Crypto #Onchain #tradingStrategy