Zach Anderson
Mar 13, 2026 17:07
New analysis reveals backtests utilizing revised on-chain knowledge produce deceptive outcomes. Level-in-time metrics reveal considerably worse real-world efficiency.
That worthwhile buying and selling technique you backtested? It most likely would not have labored in actual time. Glassnode’s newest analysis demonstrates how retroactively revised on-chain knowledge creates a harmful phantasm of profitability that evaporates when examined in opposition to data merchants truly had entry to.
The analytics agency ran similar backtests on a easy BTC trade steadiness technique—one utilizing normal historic knowledge, one other utilizing immutable point-in-time (PiT) metrics. Similar sign logic, identical parameters, identical 0.1% buying and selling charges. The outcomes diverged dramatically.
The Hidden Downside with On-Chain Knowledge
Metrics like trade balances aren’t static. They get revised as handle clustering improves and entity labeling updates. That Binance BTC steadiness determine you are for January 15, 2024 might not match what was truly printed on that date.
Whenever you backtest in opposition to revised knowledge, you are buying and selling on data that did not exist when choices would have been made. This look-ahead bias is especially extreme for metrics depending on entity identification—precisely the sort of knowledge many merchants depend on for trade movement evaluation.
Glassnode’s take a look at technique was simple: go lengthy when the 5-day shifting common of Binance’s BTC steadiness drops under the 14-day common (sustained outflows), exit when it crosses again above (outflows reversing). Working from January 2024 via March 2026 with $1,000 preliminary capital, the usual backtest confirmed efficiency roughly matching buy-and-hold.
The PiT model advised a unique story. Whereas each methods tracked equally via a lot of 2024, the immutable knowledge model missed the sturdy November 2024 and March 2025 rallies that the revised-data backtest captured. Cumulative efficiency ended up “significantly decrease,” in line with Glassnode.
Why This Issues for Quant Merchants
The implications lengthen past this single technique. Any backtest counting on knowledge topic to revision—trade balances, entity-tagged flows, even buying and selling volumes from exchanges that report with delays—faces the identical contamination threat.
This aligns with broader issues in quantitative finance about knowledge high quality. Analysis from various knowledge suppliers reveals PiT methodology prevents a number of bias varieties: look-ahead bias from utilizing future revisions, survivorship bias from datasets that exclude failed entities, and hindsight bias from restated figures.
For crypto particularly, the place on-chain analytics corporations repeatedly refine their entity labeling and clustering algorithms, the revision drawback compounds. A pockets recognized as belonging to Binance at present may not have been tagged accurately two years in the past when your backtest assumes you traded on that sign.
The Sensible Repair
Glassnode now affords PiT variants for all metrics via their Skilled tier. These append-only datasets lock in every knowledge level because it was initially computed—no retroactive modifications.
The tradeoff is actual: your backtests will doubtless look worse. However they will mirror what would have truly occurred. For merchants allocating actual capital based mostly on quantitative alerts, that accuracy hole between a flattering backtest and disappointing reside efficiency could be costly.
Earlier than deploying any technique constructed on on-chain metrics, the query is not whether or not the backtest seems worthwhile—it is whether or not you examined in opposition to the information you’ll have truly seen.
Picture supply: Shutterstock