What can go wrong, how Soomario manages risk across the strategy suite, and how to size your own exposure responsibly.
No algorithmic trading system wins every trade. No backtest perfectly predicts the future. No amount of sophistication eliminates the possibility of loss. If anyone tells you otherwise, they're lying to you.
What a well-designed system does is manage risk — controlling the size of losses, limiting exposure to any single position, and ensuring that the inevitable losing trades don't wipe out the gains from winning trades. That's what separates sustainable strategies from gambling.
A drawdown is the peak-to-trough decline in your account before a new high is reached. Every strategy experiences drawdowns — the question is how large and how long. Max Pain's six-year backtest showed an 11.8% maximum drawdown across 1,184 days of trading; Farms operates under a portfolio-level drawdown halt at 12% with auto-resume; Zones uses graduated DCA with a catastrophic stop outside the deepest level. A 15% drawdown means a $10,000 account temporarily drops to $8,500 before recovering. Live drawdowns can run 1.5×–2× backtested drawdowns once slippage, regime shifts, and out-of-sample volatility are accounted for. See How to Read a Backtest for how to weight backtested drawdown numbers honestly.
Most Soomario vault strategies use modest leverage. Aphelion and Elite cap at 2× isolated; Max Pain runs 2×–3× tiered per coin. Farms uses modest grid leverage with hard stops 1.5% beyond the lowest grid level. Leverage multiplies both gains and losses identically. A 5% adverse move at 5× leverage is a 25% loss on the position's margin. The leverage caps are deliberately tuned so that normal volatility clusters do not trigger liquidation. Higher leverage would mean tighter stop bands, which would mean more false stops in normal market noise — and false stops are more expensive than wider drawdowns. The Accumulator subscription product uses no leverage at all, executed by the user on their own broker.
Strategies that work in one market environment may underperform in another. A momentum strategy thrives in trending markets but suffers in choppy, range-bound conditions. A mean-reversion strategy thrives in choppy conditions but underperforms in strong trends. A range-farming strategy earns from oscillation around an anchor — when assets enter persistent one-direction trends, the anchor stops being meaningful and the strategy rotates out. Soomario mitigates regime risk by running strategies with different shapes — momentum (Elite), mean reversion (Max Pain, Zones), accumulator (Aphelion, Accumulator), range farming (Farms), volatility harvesting (Premia) — but no single product is immune to its own regime risk. See Mean Reversion vs Momentum.
Backtested results are calculated on historical data. Live trading faces challenges that backtests don't fully capture: slippage, liquidity gaps in extreme volatility, exchange downtime, network latency, funding costs that drift over time. Soomario publishes both backtested and live vault metrics so you can compare. Live performance underperforming backtest by 20–40% is normal; underperforming by more than that is a signal to dig deeper.
This is the most consequential design choice in the platform. Every leveraged vault strategy places its stop-loss orders directly on the exchange order book at the moment of entry — not as an alert that triggers a bot to close the position, but as an actual reduce-only order resting on the matching engine. If a stop has to fire, it fires whether the bot is alive, reachable, rate-limited, or restarting. It fires at a price the strategy chose, against the exchange's full liquidity, with no dependence on bot uptime. Bot-side monitoring runs as a redundancy on top, not as the primary defence.
Max Pain runs an independent 45-second stop-loss safety loop in addition to the TradingView signal layer. Farms places hard stops 1.5% beyond the lowest grid level on every active grid, with ATR trailing on the seed position. Zones uses a portfolio-level circuit breaker plus a catastrophic stop outside the deepest DCA level. The pattern is consistent: stops live where they don't depend on the strategy bot being awake.
No single trade risks the entire account. Strategies allocate small fractions of equity per trade — Accumulator's signal-multiplier system caps at 5× the user's daily DCA budget, Zones' DCA layers are graduated so deeper entries get bigger allocations but the total exposure is bounded, Farms runs at most five concurrent grids each sized to a fraction of vault equity. A string of losing trades cannot eliminate the account.
The platform's vault products are decoupled by design — they use different signal categories, target different inefficiencies, and reach their drawdowns at different times. A range-farming strategy (Farms) earning steadily in choppy markets can be flat in trending regimes that punish it; a mean-reversion strategy (Max Pain, Zones) can prosper in chop while a momentum strategy (Elite) waits. This is structural diversification, not just asset diversification: spreading capital across decoupled strategy shapes reduces synchronised drawdown risk.
A subtle but important risk in automated trading is state drift between the strategy's internal record and the exchange's actual state. An API failure could be misinterpreted as "no positions exist" and trigger a cascade of erroneous closes. Soomario's strategies require two-stage confirmation for any state-changing close action: a fresh, force-refreshed read from the exchange has to confirm the close decision before the action is taken. If the exchange contradicts the close decision, the close is aborted entirely. This guard is invisible during normal operation. It exists for the day when something unexpected happens.
Every product page publishes risk metrics: max drawdown, Sharpe ratio, Sortino ratio, profit factor, and win rate. Vault addresses are public, so anyone can verify positions and trade history on Hyperliquid in real time. There is no back-office, no LP statement, no quarterly attestation — the system is its own audit.
Rule 1: Never invest more than you can afford to lose entirely. Algo trading is not a savings account. Even with mechanical stops, the worst-case loss is the entire allocated capital.
Start small. The Accumulator at $7/month with no leverage is the lowest-risk way to experience signal-driven trading. If you move to vault products, start with a small deposit and increase only after you have watched the strategy through at least one full drawdown-and-recovery cycle.
Match the product to your risk tolerance. Accumulator (no leverage, executed by you) is fundamentally different from a leveraged perpetuals vault. Read each product's honest-limits section before depositing.
Size for max drawdown, not for return. If a strategy has a backtested max drawdown of 15%, assume live drawdowns can be 22–30% in the worst case. Could you tolerate that without panic-withdrawing? If not, size the deposit smaller, or pick a strategy with a tighter drawdown profile.
Diversify across strategy shapes, not just dollar-amounts. Splitting capital evenly between two long-bias strategies that share the same drawdown drivers is barely diversification. Splitting between a long-bias accumulator (Aphelion), a non-directional range farmer (Farms), and a mean-reversion vault (Zones or Max Pain) is structural diversification.