Four removed in September. One of them was up on the month.

September return: 10.62%

The Index stood at 122.33 at the 1 September observation and 135.32 at the 1 October observation. That is a 10.62% monthly return and 35.32% since inception on 1 July.

July returned 11.09% and August 10.11%. All three monthly returns were positive; this period does not show how the Index behaves through a falling month.

Drawdowns and recovery

The largest drawdown on the record remains 11.03%, from 16 to 22 August. The Index passed its previous peak at the 5 September observation, fourteen daily observations after the trough.

Within September, the deepest drawdown was 3.47%, from 130.81 on 7 September to 126.26 on 14 September. The Index reached 136.51 on 27 September and ended the month at 135.32.

Four constituents were removed

Three constituents breached the loss limit set at the 1 September rebalance, on 4, 18 and 19 September. A fourth breached the trailing drawdown limit on 14 September while still up on the month. Each was removed at that observation’s close and contributed no return from the next interval. None was reinstated during September.

The trailing drawdown rule measures the fall from a constituent’s high-water mark. A constituent can therefore be removed while its return for the month remains positive. August had one such removal as well.

Of the 30 strategies in the September basket, 25 ended the month above their starting level and five below, according to the constituent record. The rule removed ten constituents in August and four in September.

Correlation varied by month

Across 94 daily observations from 1 July to 2 October, the Index’s daily-return correlation was −0.09 to bitcoin and −0.16 to ether.

The bitcoin correlation was 0.28 in July, −0.70 in August and 0.62 in September. September’s correlation to ether was 0.64. These monthly values differ substantially from the full-record figure; the record is still short.

The 1 September note used 15-minute prices matched to actual capture times for an earlier window. Its reported 0.48 is therefore not directly comparable with the daily-candle calculations here.

Limits of this record

Three rising months do not establish manager skill. The monthly return alone cannot separate trading decisions from market conditions.

Drawdowns use one scheduled observation per day. Intraday losses may have been larger.

The 25 up and five down figures are counts, not published constituent returns.

Four removals in September do not predict the number in October. The record does not establish constituent capacity, liquidity, accessibility or implementation feasibility.

A falling month will provide a more useful test of drawdown control and the removal rule.

Method and sources

OSIQ 30 uses equal-weight Index construction and flow-adjusted, time-weighted whole-account constituent returns. Monthly returns compare the first scheduled observation of one month with the first of the next. Drawdowns use the deepest peak-to-trough move in the published daily observation series.

Correlation is the Pearson correlation of daily simple returns, comparing the Index observation at 00:05 UTC with the open of the same day’s Hyperliquid daily candle for BTC and ETH perpetuals. Correlation windows run through 2 October 2026. Method version: Tracking v3 whole account · Selection V8.

There were no corrections in September. The public record lists one correction, dated 25 August 2026. Constituent identities and limit thresholds are not published.

Index inclusion is not an O.P.E.N. Profile, a human-reviewed Strategy Passport assessment, an investment recommendation, or an investability finding. OSIQ 30 is a public research benchmark and is not currently investable.