Trading Concepts

This page explains the core ideas behind AurexTrade in plain language. No prior trading experience needed.

Paper Trading vs Live Trading

When you start the bot for the first time, it trades with virtual money in your OANDA practice account. This is called paper trading — it works exactly like real trading, but nothing you do costs real money.

Mode	Money	Risk	Purpose
Paper (practice)	Virtual — provided free by OANDA	None	Learn how the bot works, test strategies
Live	Real money from your funded account	Real financial risk	Actual trading for profit/loss

You should stay on paper trading until you're confident in your strategy and settings. There's no time limit and no pressure to switch.

Going live requires two confirmations

AurexTrade has a safety gate: switching to live trading requires you to explicitly confirm twice. You can never accidentally trade real money.

What Is a Trading Strategy?

A trading strategy is a set of rules that tells you when to buy and when to sell. Instead of making decisions based on gut feeling, the strategy follows the same rules every time — no emotions, no second-guessing.

Think of it like a recipe: given specific ingredients (market data), follow these steps, and you get a specific output (a buy or sell decision).

The Strategies

AurexTrade offers two strategies, both grid strategies built for gold (XAU_USD). They don't try to predict direction — they place orders around the current price and manage risk through session and daily limits.

Ciby Sliding Grid (Direction-Neutral, Sliding Window)

The idea: A variation on the Ciby hedged grid that follows price with a small, fixed window of levels instead of leaving every level open forever. It opens a hedged pair (one buy AND one sell) at each level, but keeps only a handful active at a time — as price trends, levels far behind are closed to free up margin, while the anchor pair is never closed and its winning side rides the whole move. That anchor leg is where the real profit comes from.

How it works:

When the bot starts a session, it takes the current price as the anchor and opens a hedged pair there. It then sets the first level a wider anchor gap away (above and below), and every level beyond that one grid spacing apart. For example, with an anchor of 4100, an anchor gap of 15 and a spacing of 10, the levels sit at …4085, 4100 (anchor), 4115, 4125, 4135…

At each level the sell rests at the level price and the buy rests a small offset above it (typically $0.90) to work around the spread.
Each side carries a stop-loss just past the next level in its losing direction — a buy is stopped below, a sell above — so a stopped-out leg loses only about one grid gap.
Orders rest at their exact price until the market reaches them (no early fills), so the grid is laid down precisely even in fast-moving markets.
Position size is smaller at the anchor and larger at every other level.

The sliding window (margin management):

The bot caps how many levels are active at once — by default 2 levels on the side price is trending into and 1 on the trailing side (the anchor is exempt and always counts as neither). As price climbs and a new level opens beyond the cap, the trailing level nearest the anchor is closed, banking its result and freeing margin for the level ahead.

For example, anchored at 4100 with price rising:

Price reaches 4115, then 4125 → both active (2 above the anchor).
Price reaches 4135 → opening it would make 3 above, so 4115 is closed → active above = 4125, 4135.
Price reaches 4145 → 4125 is closed → active above = 4135, 4145.

The anchor's pair at 4100 stays open the entire time. Because price kept rising, the anchor's buy leg is deep in profit — that is the position the strategy is really riding. Closing the trailing levels isn't a loss; those hedged pairs are closed at or near break-even, and the point is simply to stay in the game without running out of margin. If price turns and trends down instead, the window flips: 2 active levels below the anchor and 1 above.

Re-completing a pair: If price leaves a level (still within the active window) and later returns after one side was stopped out, the bot re-places only the missing side — completing the pair again without stacking duplicates. Levels that were deliberately closed for margin are not reopened.

Session management: Same safety net as the hedged grid — a session profit target and session loss limit (close everything and restart fresh), plus a daily loss limit that stops trading until the next day. All three limits measure your true profit or loss — what you have actually banked from closed trades plus the current floating value of any positions still open. This matters for a hedged grid, where a stop-loss on one side is usually paired with an open winning position on the other: the limit only trips when your real, net position breaches it, not when one half dips on its own.

When it works best: Volatile instruments that trend (gold/XAU_USD is ideal). A sustained directional move lets the anchor leg run while the window keeps margin under control.

When it struggles: Tight whipsaw that repeatedly stops out both sides of pairs near the same levels.

Analogy: It's like a moving walkway with a fixed number of footholds. You plant your anchor foot and let it carry you the whole way; as you stride forward, the foothold furthest behind folds away so you're never overstretched — but you keep moving in whichever direction the walkway is taking you.

Ciby Hedged Doubling Grid (Breakout Capture)

The idea: Do nothing in sideways markets, capture big directional moves, and never bleed from whipsaw. Instead of using stop losses that get picked off in choppy conditions, this strategy uses hedged pairs with no stop loss and only profits from doubled positions at outer grid levels.

How it works:

When the bot starts, it places 4 grid levels around the current price (2 above, 2 below). At each level, it opens a hedged pair (one buy AND one sell) with no stop loss. Hedged pairs always net to zero P&L — they exist to mark that price has visited a level.

The real action happens at the outer levels (the 2nd level from the anchor):

When price drops to the outer-below level: an extra buy is placed (betting on a bounce back up)
When price rises to the outer-above level: an extra sell is placed (betting on a reversal back down)

These extra "doubled" positions have a trailing stop that locks in profit once the price moves favourably by one grid spacing.

Take profit: Every trade has an automatic broker-side take-profit set at 2 grid spacings from entry. Long trades close when price rises 2 spacings; short trades close when price drops 2 spacings. This locks in profit per-leg before price can reverse and neutralise gains.

Protection mechanisms:

Auto take-profit on every trade — each leg closes automatically when price moves 2 spacings in the profitable direction, banking gains before price reverses
No stop loss on hedged pairs — eliminates all whipsaw bleeding
Trailing stop on doubled position — captures breakout profit, limits giveback
Session loss limit — circuit breaker if the doubled position goes against you
Whipsaw detection — if the same level re-triggers too many times, the session pauses automatically

When it works best: Markets that range quietly, then break out in one direction. Gold (XAU/USD) during news events or session opens is ideal.

When it struggles: Sustained adverse moves immediately after doubling (the doubled position loses). However, loss is bounded by the session loss limit.

Analogy: It's like setting a net at the edge of a fish pond. You don't catch anything while the fish swim in circles in the middle (zero cost). But when they make a break for the edge, you catch them.

Simple Grid (Direction-Neutral)

The idea: Place a grid of price levels above and below the current price, and let the market reveal which way it wants to go. Losing positions get stopped out; winning positions accumulate.

How it works:

Imagine drawing horizontal lines on a price chart at regular intervals (e.g., every 10 points) above and below where the price is right now. These are your grid levels.

When price crosses a level upward — buy (the market is breaking out higher)
When price crosses a level downward — sell (the market is breaking out lower)

Each position gets a wide stop-loss (e.g., 30 points) to give it room to breathe. Positions that go the wrong way get stopped out. Positions that go the right way survive and accumulate — gradually shifting your total exposure in the direction the market is actually moving.

When it works best: Choppy or slowly trending markets. Gold (XAU/USD) is a natural fit due to its tendency to oscillate within ranges before breaking out.

When it struggles: Fast, strong directional moves where price blows through multiple grid levels at once, triggering stops on several positions in rapid succession.

Key risk: Multiple stops can trigger in quick succession during strong trends. The wide stop distance and max-levels cap are your defences against this.

Analogy: It's like casting a net in both directions and seeing which side catches fish. You don't need to know which way the fish are swimming — the net does the work.

Which Strategy Should I Use?

Market Condition	Best Strategy
Sustained trends, with margin kept under control	Ciby Sliding Grid
Quiet ranges followed by breakouts	Ciby Hedged Doubling Grid
Not sure	Test with a backtest and compare results

In practice, no one knows in advance what the market will do. That's why testing matters — you can see which strategy worked better on recent data.

What Is the Risk Engine?

The risk engine is a safety system that sits between the strategy and your money. Even if the strategy says "buy", the risk engine can block the trade if it would be unsafe.

Rules the risk engine enforces:

Rule	What it prevents
Kill switch	Halts all trading immediately (emergency stop)
Stop-loss required	Every trade must have a maximum loss limit
Max drawdown	Stops trading if total losses get too large
Consecutive losses	Pauses after several losses in a row
Position size limit	Prevents betting too much at once
Daily loss limit	Stops trading for the day if losses exceed a threshold
Trade frequency	Prevents excessive trading (overtrading)

Think of it as a co-pilot that can override the autopilot when things look dangerous.

Understanding Risk Settings

When you start the bot, you'll see a "Risk Settings" section. Here's what each setting means in plain English:

Kill Switch: An emergency stop button. When activated, the bot immediately stops placing any new trades. Use this if something feels wrong and you want everything to halt right now.
Risk Per Trade (default: 2%): How much of your balance you're willing to risk on a single trade. At 2%, if your balance is $100,000, the most you could lose on one trade is $2,000. Lower = safer but slower growth. Higher = more aggressive.
Max Position (default: 10 units): The largest trade size the bot can place at once. Think of it as a cap on how big any single bet can be.
Max Daily Loss (default: $500): If the bot loses this much in a single day, it stops trading for the rest of that day. Prevents a bad day from becoming a terrible day.
Max Drawdown (default: 20%): If your balance drops more than 20% from its highest point, the bot stops entirely. This is your "worst case" safety net.
Max Consecutive Losses (default: 5): If the bot loses 5 trades in a row, it pauses. This gives the market time to settle before trying again.
Max Trades Per Day (default: 10): Prevents the bot from trading too often. More trades = more fees, and sometimes doing less is better.
Require Stop-Loss (default: on): Every trade must have a maximum loss limit set before it's placed. If the market moves against you, the trade closes automatically at that limit. This should almost always stay on.

Start with the defaults

The default risk settings are conservative and designed for safe practice trading. You don't need to change them until you understand what each one does and why you'd want a different value.

The 3-Step Validation Workflow

Before you trust a strategy with real money, you need to validate it. AurexTrade provides three tools for this:

Step 1: Backtest

What it does: Replays past market data through your strategy and records what would have happened.

Why it matters: If a strategy doesn't work on past data, it almost certainly won't work on future data either.

Limitation: Just because it worked in the past doesn't guarantee future success.

Step 2: Parameter Sweep

What it does: Tests every combination of strategy settings (e.g., different grid spacings and anchor gaps) and ranks them by performance.

Why it matters: The same strategy can be profitable with one set of parameters and unprofitable with another. A sweep finds the best settings systematically.

Limitation: The "best" settings might be overfit — they could be tailored to quirks of that specific historical period.

Step 3: Walk-Forward Validation

What it does: Splits the data into alternating "learning" and "exam" periods. Finds the best parameters on the learning period, then tests them on the exam period (data the system has never seen).

Why it matters: This is the closest thing to testing on future data without actually waiting. If settings work on the exam period, they're more likely to be genuinely robust.

Analogy: It's like studying for an exam using practice questions (learning period), then taking a real exam with different questions (exam period). If you pass the real exam, you probably actually learned the material — not just memorised answers.

What Happens When You Start the Bot?

Once you've validated a strategy, you can run the trading bot. Here's what it does each cycle (every 60 seconds by default):

Fetches the latest price data from OANDA
Runs the strategy to see if there's a buy or sell signal
Checks the risk engine — is it safe to trade?
Places the order if approved (or skips if rejected)
Records everything — every decision is logged

The bot runs continuously until you stop it or the kill switch activates.