AI Strategy Builder: Turn Your Trading Rules Into a Strategy
How an AI strategy builder turns your own written trading rules into a backtested strategy without coding: the flow, a rule template, costs, limits and Pine Script.
Many traders have a clear set of rules in their head or in a notebook, but no way to turn them into something a computer can test. Learning to code is one route. An AI strategy builder is another: you describe your own rules in plain words, and the tool translates them into a strategy and runs a single backtest. This guide explains what that process looks like in the WeCore AI Strategy Builder, how it relates to Pine Script, how to write rules clearly enough to be built, and how to read the resulting backtest without fooling yourself.
One point first. The builder does not come up with ideas. It translates what you wrote, and every number in the strategy is yours.
Pine Script in one minute
Pine Script is TradingView's own programming language. A script that begins with the `strategy()` declaration is a strategy: it simulates hypothetical orders on historical and real-time bars, and TradingView shows the outcome in the Strategy Tester (a Metrics tab and a list of trades). According to the Pine Script strategies documentation, on historical bars the broker emulator fills orders after a bar closes. The rest of the language is described in the Pine Script documentation.
That is useful, but it means you need to write and debug code. Typical stumbling blocks are order logic, position sizing and getting exits to behave. If you already write Pine, our guide on TradingView alerts and webhooks covers sending its signals onward. If you do not, a builder is an alternative path to the same destination: a rule-based strategy that can be tested.
What the WeCore AI Strategy Builder does
You find it inside the WeCore app under Tools -> Strategy Builder. It needs an active Start, Prop or Fleet plan or a WeAlgo subscription. The flow is:
- Describe your rules in plain words, or paste your own Pine Script.
- Pay for the session (card, credit or a code).
- Answer a few short questions. If a value is missing, the builder asks instead of guessing. Short informal answers are fine, and it does the sizing arithmetic for you.
- Review and confirm the specification card. This is the key checkpoint: a written summary of exactly what will be built.
- Generation and one backtest.
- Save the strategy, or finish without saving. Either choice locks the session as read-only.
If a rule cannot be built exactly with the available building blocks, the builder says so, explains an alternative, and builds that only after you confirm. The only settings it may fill in without your words are neutral operational ones, such as an end-of-day flatten time, trading days or price source, and they appear on the specification card for you to review.
What it will not do
- It is not advice. It never recommends, picks or "improves" parameters for you.
- It never predicts how a strategy will perform.
- It does not invent a strategy from nothing, and it cannot reproduce WeCore's own ready-made strategies.
- It does not interpret a backtest as a recommendation.
What it can build
The supported instruments are futures such as MNQ/NQ, MES/ES, MYM/YM, M2K/RTY, MGC/GC, MCL/CL, SIL/SI and M6E/6E, on 1, 5 and 10-minute charts. It offers around 50 indicators (moving averages, RSI, MACD, ATR, Bollinger Bands, VWAP, session ranges and more), common templates such as opening range breakout, moving average cross, mean reversion and time-of-day rules, and exits and risk settings: a mandatory stop, up to three targets, breakeven, trailing, time exit, end-of-day flatten, risk-based sizing and daily limits.
How to write a rule spec the builder can use
Vague input produces many clarifying questions, or a strategy that does not match what you had in mind. Before you open the builder, write the rules down. The template below covers what almost every strategy needs.
| Part | What to state | Example of the form (not a suggestion) |
|---|---|---|
| Instrument and timeframe | One symbol, one chart interval | "MNQ, 5-minute" |
| Session | Days and hours the strategy may trade | "09:30-11:30 New York, weekdays" |
| Entry | Exact trigger, direction, order type | "Long when the 20 EMA crosses above the 50 EMA, limit order at the close" |
| Stop | A fixed distance or structure, always defined | "N ticks below entry" |
| Targets / exit | One or more targets, time exit, trailing | "Single target at M times the risk" |
| Sizing | Fixed contracts or risk per trade | "1 contract" or "X dollars risk per trade" |
| Daily limits | Max trades, max daily loss | "Stop trading for the day after N trades" |
| Flatten time | When open positions must be closed | "Flat by a stated time" |
The examples show only the shape of each answer; the values (N, M, X) must come from your own thinking. Three habits help:
- Write conditions you can check on a chart. "The market looks weak" cannot be tested. "Close below the 50 EMA" can.
- Decide what happens in awkward cases. What if a signal fires while you are in a position? What if the stop and target are both touched in one bar?
- One idea per strategy. A fix session cannot change the symbol, timeframe or core entry logic, so settle those before paying.
Costs and limits, plainly
- Included every month: Start 1, Prop 2, Fleet 5, WeAlgo 5 new-strategy builds (with both a plan and WeAlgo, the higher number; unused builds do not carry over; 1 for the whole free trial).
- Each extra new strategy: $10, paid before any AI work. It includes up to 6 clarification rounds, one specification confirmation, generation with up to 3 automatic fixes, one backtest, and saving. The session stays open for 48 hours.
- Fix session: $5. Up to 3 change rounds and one backtest, producing a new version of the same strategy. You can change parameters and existing rules, but not the symbol, timeframe or core entry logic. That needs a new $10 build.
- A plan or WeAlgo subscription is required. A code or credit never replaces it.
- Saved strategies are unlimited, and a saved version never changes; a fix creates the next version.
- If a session fails on WeCore's side, you automatically receive a credit of the same type.
Plans start at $49 per month ($39.20 per month, billed yearly) and are listed in the pricing section.
Reading the backtest honestly
The builder runs one backtest on stored historical data, with commission and no slippage, filling at bar close, stop or target prices. Treat the output as a hypothetical description of the past, nothing more. Some practical habits:
- Remember what is missing. No slippage means every fill is better than most real fills. Our article on backtesting vs live trading lists the other gaps, from latency to data quality.
- Check the number of trades. A handful of trades says very little about anything.
- Look at the whole period, not one headline figure. Ask whether the result depends on one short stretch of time.
- Be wary of repeated tweaking. Each time you adjust a number to improve the past result, you move toward overfitting. Decide your rules for reasons you can explain, not because a number looked better.
- Verify outside the backtest. Use simulation first; see simulation mode before going live.
Past results are no indicator of future results. The builder will not tell you a strategy is good, and neither should you assume it from a chart.
Where a saved strategy runs
A saved strategy runs only on your own trading server, never on WeCore's central system. You can build and save without one, but you cannot run it until you have one. After saving, the strategy appears under Strategies -> My strategies and as a custom strategy in the WeAlgo tab, where you assign it to accounts. Running it needs an active WeAlgo / customer-server subscription. Read more in running strategies on your own server, and see the ready-made WeAlgo strategies page if you would rather look at existing options. Position sizing and risk limits remain yours to set; position sizing methods is a good primer.
Is this the right route for you?
If you already trade by written rules and want them tested without learning to code, a builder fits. If you enjoy writing Pine Script, you may prefer to stay in TradingView. Either way, a clear rule spec comes first. When you are ready, you can create an account and look at the Strategy Builder inside the app at your own pace.
FAQ
Do I need to know Pine Script to use an AI strategy builder?
No. You describe your rules in plain words. You can also paste your own Pine Script if you already have it.
Will the builder suggest a strategy or the best settings?
No. It only translates your own rules, asks for any missing values, and never recommends or optimizes parameters for you. You decide every number.
How much does it cost?
Each subscription includes new-strategy builds every month (Start 1, Prop 2, Fleet 5, WeAlgo 5). An extra new strategy session is $10 and a fix session is $5, paid before any AI work begins. You need an active Start, Prop or Fleet plan or a WeAlgo subscription.
Does the backtest tell me the strategy will work live?
No. It is hypothetical, uses commission but no slippage, and past results are no indicator of future results.
Where do saved strategies run?
On your own trading server, not on WeCore's central system. Without a server you can build and save but not run.
This article is general information, not investment advice, and makes no promise of results. Backtests are hypothetical, and futures trading carries a high risk of loss.