Strategy, bot, and execution
In Tradee, a strategy is a saved rule, a bot binds that rule to instruments, and an execution is one run of the bot with its own mode and modeled funding.
A strategy is a saved rule. A bot binds one or more copies of that rule to instruments, timeframes, and shared risk settings. An execution is one run of that bot, with its own mode and its own modeled funding. Saving the rule, or saving the bot, does not start the run.
Three jobs
People use "strategy" and "bot" for the whole system. In Tradee the words stay narrower, because each object can change without the others.
A strategy is versioned text of the rule: entry, exit, and any stop or target logic on the builder. You can backtest that definition on historical candles. The backtest is a separate job from a live run. The strategy row itself is still only the rule.
A bot is the authoring object that makes the rule operational. It picks a reference venue, which is an exchange and an asset type used for market data and reference prices. It does not store an exchange login. Under that venue, each instance pins a strategy version, a ticker, a timeframe, a signal mode, and a capital weight. The same strategy version can appear on more than one instance, on different tickers or different timeframes. Sizing and risk limits sit on the bot, with optional per-instance overrides.
A bot execution is what Execution starts. It points at a bot version and adds two choices the bot config does not version: mode, and funding. Mode is paper or output. Funding is a standalone modeled budget, or membership in a pool that shares one modeled capital envelope by weight. Editing mode or funding does not create a new bot version. A freshly saved bot has no execution until you start one. The learn path says the same thing in the bot step: creating a bot does not start trading.
One plan, two runs
The figures in this article's images are a worked example, not a recorded backtest.
That split is the useful part. You can change the shared plan once and still run paper and output together. You can also run two paper executions with different modeled budgets. Each execution has its own positions and its own funding. They do not share a fill, because Tradee does not send orders.
Output mode is not on every plan. Free accounts can run paper within plan limits. Live API outputs require Hobbyist or Pro, and a connection you assign. Both modes still stop at modeled state. What is automated trading? separates the signal from routing, exchange execution, and custody. Tradee stays on the signal side of that line.
Where a pool fits
A paper pool is not a third object beside the bot. It is one funding shape for executions. Members split a fixed modeled envelope by weight. Adaptive budget can add realized modeled results since a baseline. The envelope is not an exchange wallet, and a pool is not a broker demo account.
Paper trading versus backtesting covers what a historical test and a paper run each prove. This page only places them: the backtest reads the strategy or the bot against stored candles; the execution is the ongoing run.
What you can change without starting over
Use the object that matches the change.
- Rewrite the rule, or promote an older strategy version, when the entry or exit logic is wrong. Instances pin a version, so an old bot keeps the pin until you update it.
- Add or remove an instance, or change its ticker, timeframe, weight, or signal mode, when the same rule should watch a different market. That is a new bot version.
- Change mode or funding on Execution when the rule is fine and you want another run, another budget, or a pool slice. That does not bump the bot version.
Instance weights divide the bot's modeled slice. They do not, by themselves, make two instances take the same dollar risk. Size and stops still decide the loss if a position stops out. That question belongs with position sizing, not with the object names.
A practical order
- Write the strategy and name the assumptions: market, timeframe, entry, exit, and protection.
- Backtest that version before you trust it as a pin.
- Create a bot, add the instances, and set sizing and risk you can explain.
- Start an execution. Use paper first. Add an output execution only when a receiver is ready to consume events.
- Keep the three names in your notes. A result from an execution is not a property of the strategy file alone.
FAQ about bot execution
Can one strategy run on two markets?
Yes. Put two instances in one bot, or use two bots. Each instance pins the strategy version to its own ticker and timeframe. The strategy definition stays shared. The positions belong to the executions.
Does stopping an execution delete the bot?
No. Stopping changes the run. The bot version and the strategy version remain. A stopped execution can keep its funding and its label so you can start it again.
Methodology and sources
The object split follows Tradee's current bot and execution model: versioned strategies, versioned bots, and executions that carry paper or output plus standalone or pool funding. The counts 0, 1, and 2 are constructed so the panels match the example. They are not a measured account. This article is educational and does not provide personalized investment or trading advice.
This article is for education. It is not financial advice. Trading can lose money. Past results do not guarantee future results.