A backtest runs a rule over past data. That assumes the past data exists. For a market that opened this year, it does not, and no amount of care in the method can make up for it.
The problem in plain terms
Stock and index perps on Hyperliquid are new. So are the tokenized stocks on Robinhood Chain, and so is nearly every coin on a trending list. A perp on the S&P 500 that trades through the weekend has a price history only as long as the perp has existed. Before its first trade there is nothing to test against.
It gets thinner than that. The venue serves price history and funding history for these perps. It does not serve open-interest history. A rule that uses open interest can only be tested over the period someone happened to record it.
Why borrowing history fails
The tempting fix is to borrow. The S&P 500 has decades of data, so test the rule on the index and apply it to the perp. The trouble is that the perp is a different market. It has its own traders, its own depth, its own funding and its own hours. Weekend behavior is the part people most want to study, and the cash index has no weekend prints at all.
The same goes for borrowing from crypto perps. Funding works the same way mechanically, as funding on a stock perp explains. Whether stock perps behave like coin perps around an earnings date or a cash open is an open question, and only data from stock perps can answer it.
A short history is a small sample
Even where history exists, a few months of it holds few independent events. A year contains roughly fifty weekends. A market that opened last quarter has seen about a dozen. Any pattern drawn from so few cases is fragile, for the reasons in how many trades before you believe it.
A short history also covers few kinds of market. If the only months on record were calm, the record says nothing about a selloff. The statistics are true of the period measured and silent about the rest.
What can be done instead
The first thing is to describe instead of test. Reading what a market is doing now needs no history. A perp’s move since the cash close, its funding and its open interest are measurements, and a measurement is valid on day one.
The second is to start the record. History that was not recorded cannot be bought later, which is the point of the open interest history the venue does not serve. NoVo’s collector has recorded every stock and index perp’s open interest, funding and premium on every pass since 4 October 2026.
The third is to label honestly. Until a record is a week long, an average drawn from it covers less than a week and should say so. NoVo’s dashboards label each average with the span the record really covers.
Where NoVo shows it
The Stocks On-Chain tab on the Crypto Market Map lists every stock and index perp with its live readings and a chart of recorded open interest. The chart starts where the record starts. It makes no claim about what came before.