Safety Orders: How to Set Your First DCA Levels Correctly

Safety orders are additional buy (or sell) orders that a DCA bot places at pre-defined price steps against your entry, lowering — or raising, on shorts — the average entry price of the deal. Setting them correctly means choosing a first safety order at least the size of your start order, a price step wide enough to survive normal noise, and a martingale multiplier close to 1. You can configure all of this on a fresh bot in the GT App dashboard in under two minutes.

This article walks through the four decisions that matter on your first DCA bot — size, step, count, martingale — and the mistakes that catch beginners on both Binance and Hyperliquid. The same logic applies on both venues, since GT App launches strategies on either Binance or Hyperliquid, and Hyperliquid is a first-class venue most rival bots don’t support.

By the end you will know how to configure a bot that survives a 20-30% adverse move without a margin call, and why doubling every safety order is the fastest way to blow up a starter deposit. If you’d rather see the fields as you read, open a new bot at app.gt-protocol.io in another tab.

Table of contents

• What is a safety order?
• Sizing your first safety order
• Choosing the price step (deviation)
• How many safety orders to set
• The martingale multiplier: 1, 1.1, or trouble
• Safety orders on Hyperliquid
• Five beginner mistakes to avoid
• Frequently Asked Questions

What is a safety order?

A safety order is a limit order that a DCA bot places below your long entry (or above a short entry) at a specific price deviation, so that if the market moves against you the bot buys more at a better price and lowers the average entry of the open position. Once one or more safety orders execute, the bot cancels the current take-profit order and places a new one, recalculated from the new average — closer to the current market. The deal can then close in profit on a partial recovery, without waiting for price to return to the original entry. In GT App terms, a bot’s safety-order block has four fields: safety order amount, safety order quantity, price deviation, and martingale.

Averaging works because you are buying at progressively lower prices during a drawdown. In a canonical GT App example, a long entry at $10,000 with three safety orders at 2% steps executes at $9,800, $9,600, and $9,400. The four fills give an average entry of $9,700 — 3% below the original. A 1% take-profit target that started at $10,100 now sits at $9,797. The deal can close in profit on a bounce, not a full recovery.

Sizing your first safety order

The safety order amount cannot be smaller than your start order amount — GT App will refuse to save a bot with a smaller safety order, and there is a practical reason for the rule: a safety order smaller than the start order barely moves the average, so it burns liquidity without doing its job. The correct starting rule for a beginner is safety order amount equal to start order amount. That gives each fill the same weight in the average and keeps the math predictable.

On futures, the GT App playbook is stricter: the initial margin behind the start order should be 1–2% of your futures balance, and the safety order amount should be the same as the start order. With 20 safety orders at those sizes and no martingale, you need enough balance to place 21 identically-sized orders. That is the reason why the recommended futures workflow uses 10 to 20 safety orders and asks you to plan for the full stack up front, not order by order.

A worked example on Binance futures

Start order 10 USDT of margin (2% of a 500 USDT futures balance), safety order 10 USDT, 15 safety orders, martingale 1. Total margin committed if the entire ladder fills: 16 × 10 = 160 USDT, roughly 32% of the futures balance. The remaining 68% stays as maintenance-margin cushion. Half of the balance in the deal, half in reserve is the boundary you don’t want to cross on your first bot.

Choosing the price step (deviation)

The price deviation is how far the market has to move against you before the next safety order fires. Setting it too small (0.5%) means the whole ladder fires on normal intraday noise, and you run out of safety orders before a real drawdown finishes. Setting it too large (10%) means safety orders rarely trigger — and when they do, the position is already deep underwater. A first-bot default that works on major pairs is 2% deviation on futures, 1–2% on spot, matching the deviation used in the GT App user guide examples.

The deviation compounds down the ladder: each safety order is placed at the specified deviation from the previous safety order, not from the entry. With 2% deviation and no martingale, ten safety orders cover roughly a 20% adverse move. That range is what BTC and ETH can produce in a single week during a normal correction, so ten safety orders at 2% is a reasonable minimum floor.

What if I want a wider ladder?

Use the safety order step multiplier — GT App lets you scale each step by a fixed factor. With a 3% starting deviation and a 1.2 multiplier, the second step becomes 3.6%, the third 4.32%, and the fourth 5.18%. The first few safety orders catch normal noise; the later orders wait for real capitulation and buy the wick, not the pullback. This is the standard configuration for pairs known for sharp flash-crashes.

How many safety orders to set

The GT App recommendation is 8 to 14 safety orders for a first bot on a major pair. Fewer than 8 and you run out of ammunition on any real drawdown; more than 14 and your capital is tied up placing orders that will almost never fill, which pushes down your annualised return. Alongside the total count there is a second setting — max simultaneously active safety orders — which controls how many are actually resting on the exchange at any moment. The GT App default of 5 is a good starting value.

Placing five orders on the exchange in advance matters for a reason that is easy to overlook: during a fast drop, exchanges get overloaded and REST endpoints slow down. Orders that are already resting on the order book fill regardless. Orders the bot still has to send may miss the price. Five pre-placed safety orders give the bot a five-step cushion of resilience against exchange congestion, which is exactly when safety orders matter most.

The martingale multiplier: 1, 1.1, or trouble

Martingale on a DCA bot is the multiplier applied to each successive safety order’s size, not its price step. Martingale 1 means every safety order is the same size as the first. Martingale 1.1 means each order is 10% larger than the previous one — the 10th safety order is roughly 2.6× the first, and total capital deployed roughly doubles compared to no martingale. Martingale 2 (the classic casino martingale) means the size doubles at every step; by the 10th safety order the bot is committing 512× the first order’s capital in a single fill. On a starter deposit, this is the single fastest way to hit liquidation.

The GT App playbook gives a clear rule: pick either martingale 1 with start order 2% of balance, or martingale 1.1 with start order 1% of balance. Both configurations keep total exposure roughly equal. Martingale 1.1 is the marginally better option — the larger later orders lower the average faster during a deep drawdown — but only if you keep the start order at 1%. Setting martingale above 1.2 on a first bot is not a strategy, it is a bet that the drawdown will end before your balance does.

Safety orders on Hyperliquid

The same DCA logic applies when your bot runs on Hyperliquid. Hyperliquid is an on-chain perpetuals DEX with a fast order book, which makes it well-suited to automated strategies: your safety orders rest on-chain, fill non-custodially, and there is no centralised custodian to freeze the account. GT App connects to Hyperliquid through GT Magic — a Telegram-native onboarding that connects your wallet with one signature. You keep your keys; the bot only gets permission to trade.

Sizing and stepping rules do not change between venues. What does change is the minimum order size (Hyperliquid enforces per-symbol minimums that can differ from Binance’s), so before saving a bot on a small-cap Hyperliquid pair, double-check that your safety order amount meets that pair’s floor. If the bot refuses to save with a “minimum order size” error, raise the safety order amount rather than lowering the count. The count is what buys you drawdown coverage; the individual order size is fungible.

Five beginner mistakes to avoid

The five errors below account for almost every “my bot liquidated” support ticket on a first DCA deployment. Read them once before you save your first bot, and you will skip a painful few weeks. Each one has a fix that takes fewer than 30 seconds in the GT App UI, and none of them requires you to understand exchange internals or write any code. The pattern is always the same: an early configuration choice that looked reasonable in isolation but that stacks badly with the other three settings on the safety-order block.

• Too few safety orders (2–3). A three-step ladder covers roughly 6% of adverse movement. On BTC that is Tuesday. Set at least 8.
• Aggressive martingale (≥1.5). Doubling exposure at every step means the bot commits nearly all of your balance by the 7th or 8th safety order. Stay at 1 or 1.1.
• Deviation too small (<1%). The whole ladder burns in the first hour of a mild pullback. Use 2% on futures, 1–2% on spot.
• Stop loss placed inside the safety-order range. The stop triggers before the last few safety orders fill, converting a manageable drawdown into a realized loss. If you set a stop, place it behind the last safety order in the ladder, or leave it off — the safety orders themselves are your risk management on a DCA bot.
• Ignoring Safe Mode. New futures bots start in Safe Mode, which caps total margin across all orders at 10% of the strategy balance. If the bot refuses to save, that cap is doing its job. Reduce order sizes or the count — do not switch to Pro Mode on your first bot just to make the save button work.

Frequently Asked Questions

What is the minimum safety order size in GT App?

The safety order amount must be at least equal to the start order amount, and each exchange enforces its own per-symbol minimum notional. On Binance futures, most major pairs sit around 50 USDT. On Hyperliquid, minimums are per-symbol; check the bot creation screen for the exact floor when you pick the pair.

Do I need safety orders at all?

Not always. A short-timeframe scalping bot that closes deals quickly on 1% take-profit can run with zero safety orders on a strongly trending market. But on a DCA strategy that holds through pullbacks, safety orders are the whole mechanism — they turn a losing entry into a winning average.

What is the difference between safety order quantity and max simultaneously active safety orders?

Quantity is the total number of safety orders in the ladder — the maximum drawdown coverage. Max active is how many of them are pre-placed on the exchange order book at any moment. With 10 quantity and 5 active, the bot places the first five; as they execute, it places the next ones in turn.

Should I use martingale on a spot bot?

Martingale on spot is safer than on futures because there is no leverage and no liquidation. But the discipline is the same — 1 or 1.1 on a first bot, never higher until you have seen the strategy survive a full drawdown-and-recovery cycle.

Can I edit safety order settings while a deal is open?

Not on the active deal — the ladder is fixed once the deal opens, so the bot can honour the calculated average and take-profit. New settings apply to the next deal. This is why sizing them correctly before launch matters.

What if my safety orders never trigger?

That is the good outcome. If the market moves in your direction from entry, the take-profit fires and the deal closes in profit without any safety orders being needed. The safety-order block is insurance; you want to configure it once and never rely on it.

Do safety orders work the same way on Hyperliquid perps?

Yes. The DCA logic — averaging entry, recalculating take-profit, honouring the price step — is identical. The only differences are the per-symbol minimum order sizes and the fact that your orders rest on an on-chain order book instead of a centralised exchange’s engine.

Configure your first bot

Open app.gt-protocol.io, create a new bot on BTC/USDT, and set start order = safety order = 1% of your futures balance, 10 safety orders, 2% deviation, martingale 1.1, five simultaneously active. That is a defensible first configuration on either Binance or Hyperliquid. Run it in demo for a few days, watch how the safety orders fill during a drawdown, then decide whether to widen the deviation or add more orders. First configurations rarely stay first configurations for long — but the four decisions above are the ones that keep the deposit alive long enough to iterate.