The CFTC’s 15 September positioning snapshot carried two numbers pointing in opposite directions:
- Leveraged funds cut their net short across four regulated bitcoin futures products by 7,275 BTC-equivalent;
- Asset managers cut their net long by 4,733 BTC-equivalent over the same week.
Less short on one side, less long on the other. Both were shrinking exposure, not adding to it.
Positioning moved on the regulated futures side. What did perpetual funding do? Nothing.
Funding over the same stretch: sitting on the baseline
Binance BTCUSDT’s last six settlements:
0.010000% 0.010000% 0.010000% 0.007524% 0.004200% 0.010000%Four of six are exactly 0.010000% — not “close to,” but equal to the interest rate to the last digit. Today’s current reading at 08:07 UTC is 0.010000% as well.
That exactness is not a coincidence, it is the dead zone in the clamp formula: as long as the premium index sits between −0.04% and +0.06%, the printed rate is identically 0.01% regardless of what the premium actually is. In other words, those four periods told you nothing at all. The mechanism is in 790 periods, not one above 0.01%.
Two markets telling different stories, and only one of them pays
The difference in kind is what is worth laying out:
| CFTC positioning report | Perpetual funding | |
|---|---|---|
| What it reflects | Who holds what on regulated futures | The perpetual’s live premium to spot |
| Frequency | Weekly, reported with a lag of days | Every 8 hours |
| Coverage | CME and other regulated venues | Exchange perpetual books |
| Relation to return | No direct cash flow | It is the entire revenue of this approach |
The positioning report is structural information: it says how institutions are arranged. Funding is cash flow: it decides what you collect or pay this period. The two can diverge for long stretches, and routinely do.
So what does this news mean for a balanced book? Directly, nothing. However much net short the funds trimmed, not a cent of it lands in your account. Its value lies elsewhere: when shorts are being withdrawn on regulated futures while perpetual funding does not budge, the force moving price is not coming from the perpetual side.
This is precisely why we keep macro and institutional positioning out of the strategy — they can explain a move, and explanations do not generate cash flow.
What it means for a position
A balanced two-legged book cares about one thing: has funding turned negative. Not in the last six periods; the lowest was 0.004200%, still positive.
But the other half has to be said too: four periods pinned at the baseline means the return in those four periods was the baseline itself. At 0.01% per eight hours that annualises to 10.95%, whereas Binance BTCUSDT’s actual year to date across 790 periods is 2.103149%, annualising at 2.92% — far below the capped figure, because 209 negative periods sit in between.
Extrapolating a full year from one capped reading produces a number more than three times too high. It is the single most common misreading of funding data.
What to watch next
- The next CFTC report (the 22 September snapshot). Whether leveraged funds keep trimming shorts or start rebuilding them. It does not drive your cash flow, but it does say who is pushing this leg of the move.
- Whether the premium breaks the dead zone’s lower edge. Funding only turns genuinely negative once the premium drops under −0.04%. Until then, 0.01% carries no information.
- Do not read “institutions are covering shorts” as “funding is about to rise.” These six periods are the counterexample: positioning moved, the reading did not.
Funding drifting toward zero or turning negative is something this approach genuinely runs into — the year-by-year table on our track record has ETHUSDT at 0.787256% for 2022 and SOLUSDT at −0.941791% year to date, and we left neither out. These periods were positive; that says nothing about the next one.
Sources: the 15 September CFTC snapshot figures — leveraged funds reducing net shorts by 7,275 BTC-equivalent and asset managers reducing net longs by 4,733 BTC-equivalent across four regulated bitcoin futures products — are as reported by CryptoSlate on 2026-09-20 citing that report; Binance BTCUSDT’s last six settlements and current reading come from the public endpoints fapi/v1/fundingRate and fapi/v1/premiumIndex, read at 2026-09-21 08:07 UTC; the year-to-date 2.103149% over 790 periods and SOLUSDT’s −0.941791% come from the same endpoint, 2026-01-01 00:00 to 2026-09-21 00:00 UTC; the clamp formula and its ±0.05% bounds are from Binance’s public documentation. Annualised figures assume the current reading holds unchanged and are comparison values, not expected returns. Past performance does not indicate future results.