Aleph Edge · Risk Lab

What our bots risk together

Every bot in Aleph Edge has its own stop. This page asks the harder question: what happens to all of them at once when Bitcoin falls, and what they would have done in the worst crashes of the last six years.

Reference book: 12 templates · 61,000 USDT paper capitalData to Fri, 16 Oct 2026 04:00 UTCSame simulator as the public ghost track

Stress test: Bitcoin moves in 24 hours

The templates' open bots, with their real ladders, stops and leverage, run through a straight 24-hour move. Each coin moves by its beta to BTC.

Templates that passed or are under review (2)−771 USDT−7.71% of 10,000 USDT
All 12 templates−5,422 USDT−8.89% of 61,000 USDT
Liquidations0at leverage 1x
DCA Long Quick−26.2%
Grid Long Trail−20.7%
DCA Long Stop−15.5%
DCA Long Alts−10.9%
● DCA Long Classic−8.3%
DCA Long BTC−8.3%
● DCA Long Safe−7.1%
Grid Neutral Tight−9.2%
Grid Neutral−8.4%
DCA Short Safe+1.2%
DCA Short Classic+2.1%
Grid Short+7.1%

● passed or under review. Betas to BTC used: NEAR 2.85 · ENA 2.29 · ZEC 2.24 · 1000PEPE 2.21 · SUI 2.15 · UNI 2.10 · DOGE 1.57 · SOL 1.53 · XRP 1.50 · HYPE 1.28 · ETH 1.27 · BANK 1.11 · BNB 0.97 · AKE 0.89. Bars share one scale across all four moves.

Crash replay

Each template's bots start 7 days before the crash on that month's own coins, then live through it. Click a row to see the hours.

WindowBTC low2 under reviewAll 12All 12 worst point7 days later
Oct 2025 liquidation cascade
10 Oct 2025
−16.5%+0.7%−2.4%−2.8%−1.5%
FTX collapse
8 Nov 2022
−24.9%+3.1%−6.1%−8.6%−5.0%
LUNA / UST collapse
9 May 2022
−21.8%−27.6%−23.6%−25.5%−20.9%
19 May 2021 crash
19 May 2021
−33.1%−0.5%−9.2%−9.7%−13.1%
12 Mar 2020 (COVID) crash
12 Mar 2020
−45.2%−20.9%−13.9%−13.9%−14.7%
-35%-25%-15%-5%+5%+15%2 May8 May12 May19 May
12 May, 23:00 UTCBTC −14.7%All 12 −23.6%DCA Long Classic −28.7%
All 12, % of capitalDCA Long Classic, % of its capitalBTC, % from window startCrash window

LUNAUSDT: 1h data ends 2022-05-13T07:00:00+00:00 (bots closed at market at its last bar)

Concentration right now

Open positions, gross11,441 USDT18.8% of capital
Long share of open positions68.9%short 31.1%
Largest coin in open positionsZEC21.9% of gross
BTC-equivalent net exposure5,538 USDT9.1% of capital, average beta 1.57

Capital by template direction

LongShortNeutral grids

What changed in the app

0.3.1516 Oct 2026
  • Portfolio risk viewon the Risk page: this stress test and crash replay run on your own open bots, a "before → after" line in every bot and trade form, and alerts when one direction or one coin dominates.
  • Key safety: Binance keys that can transfer funds are refused (withdrawal keys already were); warnings for keys older than 90 days or without an IP restriction; Ed25519 keys generated inside the app, the private half never leaves the encrypted vault.
  • Foreign-trade watch: a fill on your Binance futures account that the app did not make raises an alarm, a Telegram message, and stops new entries on every bot. It ships switched off; turn it on in the Risk page.
  • TradingView alerts become paper trades through a private webhook; New trade with take-profit and stop in one form; automatic backup before every update; bot settings export and import.
  • A template warns when your pair is outside the coins it was tested on.
0.3.1415 Oct 2026
  • Telegram notifications with opt-in /status and /stop; one clear choice for what a grid does when price leaves its range; templates show the smallest budget they need.
0.3.1315 Oct 2026
  • A DCA take-profit must clear twice the round-trip fees; a warning when two futures bots share one symbol.
0.3.1213 Oct 2026
  • Paper or real money is visible on every bot; the New bot page shows each template's tested result.

How this is computed, and its limits

  • The reference book is public: the 12 templates as the ghost track runs them, 1,000 USDT per bot, leverage 1, the same simulator and fees as the template results. No user's portfolio is used.
  • The stress move is a straight line with no wicks and no funding. Real crashes wick further, so real losses can be larger.
  • Betas come from hourly returns against BTC. The ghost data starts on 6 October, so they rest on about 10 days, not 30.
  • Crash windows use each month's own top coins, so windows are not one-to-one comparable.
  • Partial data: 2020-03-12 replay: the data.binance.vision USD-M archive starts in January 2020 and most perps listed in Jan-Feb 2020, so the March 2020 universe is ranked on the 29 days of February 2020 instead of 90; only 11 USDT perps traded all of February, so dca_long_alts has 6 bots (ranks 6-11) and the all-12 capital is 57,000 USDT, not 61,000. 2022-05 replay: LUNAUSDT was a top-5 major; its futures data ends 2022-05-13 07:00 UTC (delisted), where the simulator closes those bots at market at the last bar. Stress betas: estimated on ~10 days (243 paired hours), not 30, because the ghost panel only goes back to 2026-10-06.
Paper simulation of public templates, not a track record and not advice. Generated Fri, 16 Oct 2026 09:41:36 GMT.
Offline