Long/short trend-following on USDT perpetuals, rebalanced every four hours, with an exchange-side stop on every position. A Binance Demo account and a Hyperliquid paper account have run it on simulated money since 27 September 2026, and every fill is published, losses included. No return is targeted or promised.
Live record. Since 27 September 2026, the Binance DEMO account is down 11.27% and the Hyperliquid PAPER account is down 11.50% (as of 13 Nov 2026, 12:00 UTC). Both accounts trade simulated money, not live funds. On each account, 1 of 21 completed round trips in profit. The Binance account reached a pre-set retire limit on 04 November 2026 and opens no new trades until the owner decides. The Hyperliquid account reached a pre-set retire limit on 04 November 2026 and opens no new trades until the owner decides. The review of 29 October replayed the rules over 27 September to 29 October for the Binance DEMO and Hyperliquid PAPER accounts and got the same trades, so the loss over those weeks came from the strategy in that market, not from execution. A record of about 2 months is too short to confirm or reject the strategy. Next scheduled review: 26 November 2026. Decision of 31 October and retire limits
Backtest, Jan 2021 – Sep 2026, 10 USDT perpetuals, 4-hour bars, net of fees, slippage and funding. Hypothetical results, not live trading. Sharpe ratio, from the tear sheet's monthly returns: about 2.1 in 2021–2023, about 0.8 in 2024–2026 and about 0.7 in 2025–2026, net of costs and funding; recent years have been weaker. No return is targeted or promised.
Positions: 21 opened; 21 closed, 1 in profit; 0 open, 0 in profit
Positions: 21 opened; 21 closed, 1 in profit; 0 open, 0 in profit
Positions, resting stop orders and recent fills for both accounts, laid out as each exchange reports them. Refreshed every 10 minutes. Full track record
| Symbol | Size | Value (USDT) | Entry price | Mark price | Margin | PNL (USDT) | TP / SL |
|---|---|---|---|---|---|---|---|
| No open positions | |||||||
| Symbol | Type | Side | Trigger | Amount | Reduce only |
|---|---|---|---|---|---|
| No open orders | |||||
| Time (UTC) | Symbol | Side | Price | Quantity | Realized PNL | Role |
|---|---|---|---|---|---|---|
| 05 Nov 00:00 | SOLUSDT | Buyclose | 95.690 | 10.89 | +14.05 | Taker |
| 04 Nov 16:00 | XRPUSDT | Buyclose | 1.317 | 926 | -66.58 | Taker |
| 04 Nov 16:00 | ETHUSDT | Buyclose | 2,425.6 | 0.511 | -88.06 | Taker |
| 04 Nov 16:00 | DOGEUSDT | Buyclose | 0.07722 | 15,919 | -33.49 | Taker |
| 02 Nov 20:01 | DOGEUSDT | Selladd | 0.07367 | 4,742 | — | Taker |
| 30 Oct 12:01 | DOGEUSDT | Sellopen | 0.07573 | 11,177 | — | Taker |
| 30 Oct 08:01 | SOLUSDT | Sellopen | 96.980 | 10.89 | — | Taker |
| 30 Oct 04:01 | XRPUSDT | Sellopen | 1.245 | 926 | — | Taker |
| 30 Oct 00:01 | ETHUSDT | Sellopen | 2,253.2 | 0.511 | — | Taker |
| 29 Oct 16:00 | ETHUSDT | Buyclose | 2,368.3 | 0.507 | -30.73 | Taker |
| 29 Oct 16:00 | ADAUSDT | Buyclose | 0.20630 | 5,242 | -46.65 | Taker |
| 29 Oct 08:00 | ADAUSDT | Sellopen | 0.19740 | 5,242 | — | Taker |
Binance Demo Trading: exchange matching and order handling, simulated capital. Fill prices and slippage come from the demo order book. Cross margin, leverage set to 5×. At the exposure shown, no position is close to a liquidation price. A stop order rests on the exchange for each position.
| Symbol | Size | Value (USDT) | Entry price | Mark price | Margin | PNL (USDT) | TP / SL |
|---|---|---|---|---|---|---|---|
| No open positions | |||||||
| Symbol | Type | Side | Trigger | Amount | Reduce only |
|---|---|---|---|---|---|
| No open orders | |||||
| Time (UTC) | Symbol | Side | Price | Quantity | Realized PNL | Role |
|---|---|---|---|---|---|---|
| 05 Nov 00:00 | SOLUSDT | Buyclose | 95.605 | 21.77 | +33.23 | Taker |
| 04 Nov 16:00 | XRPUSDT | Buyclose | 1.319 | 1,851 | -139.40 | Taker |
| 04 Nov 16:00 | ETHUSDT | Buyclose | 2,431.0 | 1.0227 | -185.15 | Taker |
| 04 Nov 16:00 | DOGEUSDT | Buyclose | 0.07756 | 31,830 | -76.27 | Taker |
| 02 Nov 20:01 | DOGEUSDT | Selladd | 0.07379 | 9,481 | — | Taker |
| 30 Oct 12:01 | DOGEUSDT | Sellopen | 0.07575 | 22,349 | — | Taker |
| 30 Oct 08:01 | SOLUSDT | Sellopen | 97.132 | 21.77 | — | Taker |
| 30 Oct 04:01 | XRPUSDT | Sellopen | 1.244 | 1,851 | — | Taker |
| 30 Oct 00:00 | ETHUSDT | Sellopen | 2,249.9 | 1.0227 | — | Taker |
| 29 Oct 16:00 | ETHUSDT | Buyclose | 2,369.6 | 1.0156 | -64.63 | Taker |
| 29 Oct 16:00 | ADAUSDT | Buyclose | 0.20652 | 10,485 | -94.16 | Taker |
| 29 Oct 08:00 | ADAUSDT | Sellopen | 0.19754 | 10,485 | — | Taker |
Paper account: fills simulated at live Hyperliquid prices, with hourly funding charged. Every paper fill is charged as a taker (Hyperliquid fee plus 0.02% slippage); maker fills are never simulated. Cross margin, leverage set to 5×. At the exposure shown, no position is close to a liquidation price. A stop order rests on the exchange for each position.
| Account | Opened | Round trips closed | Still open | Exits incl. trims, in profit | ||
|---|---|---|---|---|---|---|
| In profit | At a loss | In profit | At a loss | |||
| Binance DEMO | 21 | 1 | 20 | 0 | 0 | 2 of 36 |
| Hyperliquid PAPER | 21 | 1 | 20 | 0 | 0 | 2 of 37 |
Completed round trips, each position from its opening fill to its final exit with partial reductions credited to it: Binance 21, 1 in profit; Hyperliquid 21, 1 in profit. Exits including partial reductions (trims) are shown last as a secondary count; trims of one position are not independent outcomes.
Since 27 Sep, BTC has moved -17.9%, within a 29.4% range from low to high. At the backtest's 24.3% annualised volatility, one standard deviation of return over 47 days is about ±8.8%; as of 13 Nov, 12:00 UTC the accounts stand at -11.27% (Binance) and -11.50% (Hyperliquid) since start.
Trend following takes small losses when an early move fades and earns most of its return from a minority of positions that catch a sustained move. In the backtest, about four months in ten lost money. Position history by account
As of 13 Nov, 12:00 UTC, 47.3 days in: Binance -11.27% (about the 2nd percentile), Hyperliquid -11.50% (about the 1st percentile) of backtest windows of the same age. Backtest median +3.02%; 5th to 95th percentile -8.2% to +25.5%. Return since a flat start of the live rule, net of 0.07% per unit traded and Binance's historical funding, at each 4-hour close. 2,088 overlapping windows, one starting each day from 2021-01-01 to 2026-09-19; each bar uses every window long enough to reach it. Hypothetical.
Copy trading is closed. It will not open before the live record covers at least 90 days and passes its review. The cards show the account on each venue. Copy trading status and costs
Binance Demo Trading, USDⓈ-M futures
Trading on Binance Demo Trading (exchange order handling, simulated capital; fills and slippage come from the demo order book). Copy trading is closed. It will not open before the live record covers at least 90 days and passes its review.
Paper account, perpetual futures
Paper account: fills are simulated at live Hyperliquid prices and no orders are sent. Every paper fill is charged as a taker (Hyperliquid fee plus 0.02% slippage); maker fills are never simulated. Copy trading is closed. It will not open before the live record covers at least 90 days and passes its review.
Hypothetical backtest, Jan 2021 – Sep 2026, on 4-hour bars. Net of 0.07% per unit traded (fees and slippage) and of Binance's historical funding. The rule as traded until 14 October 2026; the resize rule was widened that day (journal), which changes these figures by less than their noise.
| Swarm | BTC buy-and-hold | |
|---|---|---|
| CAGR | 41.8% | 20.2% |
| Annualised volatility | 24.3% | 57.7% |
| Sharpe ratio | 1.56 | 0.61 |
| Sortino ratio | 1.97 | 0.79 |
| Max drawdown | -18.8% | -77.1% |
| Calmar ratio | 2.23 | 0.26 |
| Positive months | 57% | 54% |
| Best month | +30.3% | +43.8% |
| Worst month | -8.9% | -37.3% |
| Correlation to BTC (monthly) | 0.26 | 1.00 |
| Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2021 | +16.2% | +13.8% | -0.9% | +20.2% | +3.5% | +2.8% | +0.2% | +14.2% | -3.3% | +1.8% | +4.8% | +1.9% | +101.5% |
| 2022 | +9.3% | -2.8% | +0.7% | +2.0% | +11.2% | +5.5% | -4.6% | -5.8% | -3.2% | +10.0% | -1.2% | +9.6% | +32.5% |
| 2023 | +4.9% | -5.9% | +0.5% | -2.0% | -3.2% | +9.7% | +2.2% | +3.1% | -1.5% | +15.8% | +5.7% | +17.2% | +53.9% |
| 2024 | -4.2% | +22.6% | +6.5% | -7.7% | -7.7% | -2.4% | +4.0% | +4.2% | -2.8% | -0.7% | +30.3% | -1.5% | +39.6% |
| 2025 | -3.0% | +6.1% | -3.9% | +2.5% | +0.5% | -5.6% | +14.3% | -2.1% | -0.8% | -5.2% | +6.3% | -1.4% | +6.1% |
| 2026* | +10.4% | +1.3% | -3.4% | -4.1% | -0.8% | +14.3% | -8.9% | +12.1% | +1.7% | +22.0% |
Hypothetical results, computed on historical data with the benefit of hindsight. No capital was traded. The demo and paper accounts are reported separately above.
Calendar years 2021–2025, annualised. Hypothetical backtest 43.5%, maximum drawdown 18.8%. Bitcoin buy-and-hold 24.8%, maximum drawdown 77.1%. S&P 500 14.3%; HFRI Fund Weighted Composite 7.1%.
| Calendar-year return | 2021 | 2022 | 2023 | 2024 | 2025 | Annualised | Max drawdown |
|---|---|---|---|---|---|---|---|
| Swarm strategybacktest, after fees and funding | +101.5% | +32.5% | +53.9% | +39.6% | +6.1% | 43.5% | -18.8% |
| Bitcoinbuy and hold | +59.6% | -64.2% | +155.9% | +121.1% | -6.4% | 24.8% | -77.1% |
| S&P 500total return | +28.7% | -18.2% | +26.2% | +24.9% | +17.7% | 14.3% | n/a |
| Hedge fundsHFRI Fund Weighted Composite | +10.3% | -4.2% | +7.5% | +10.0% | +12.6% | 7.1% | n/a |
Copy trading is closed, so no profit share or vault minimum is set.
The backtest returned 6% in 2025 and its deepest drawdown was 18.8%; live results can fall below it.