Before writing a single line of code, there are three decisions that determine everything else.
BlaveClaw organizes all strategies into three types. Picking the right one first saves a lot of back-and-forth with the agent.
| Type | What it does | Backtest? | Good for |
|---|---|---|---|
| A — Signal | Trades one fixed symbol on a fixed time interval. Goes long or flat based on a rule. | Required | Most strategies. Start here. |
| B — Everything else | Screeners, grid bots, alerts, one-time order execution, anything that doesn't fit Type A or C. | None | Automation tasks, non-signal trading |
| C — Portfolio | Allocates capital across a basket of assets using weights, rebalances on a schedule. | Required | Multi-asset strategies, factor models |
The most common source of confusion is between Type A and Type C. The distinction is not about how many symbols you trade — it's about what the strategy is deciding.
| Type A | Type C | |
|---|---|---|
| The strategy decides | When to be in or out of one symbol | How to split capital across a basket of symbols |
| Signal output | 1.0 (long), 0.0 (flat), or NaN (hold) | Weight vector: [0.3, 0.2, 0.5, …] summing to 1 |
| Rebalances | When the signal changes | On a fixed schedule (daily / weekly / monthly) |
If you want to trade multiple symbols, you build multiple Type A strategies — one per symbol — and let the portfolio manager allocate capital across them. You don't need Type C for that.
| Scenario | Type | Why |
|---|---|---|
| SMA crossover on BTCUSDT 1h | A | One symbol, timing decision |
| BTC + ETH + SOL — each with its own trend signal | A × 3 | Three separate Type A strategies; the manager allocates capital across them |
| Go long BTC when Taker Intensity spikes | A | One symbol, signal-based timing |
| Weekly rebalance of 20 Taiwan stocks ranked by foreign institutional flow | C | Capital allocation across a basket on a fixed schedule |
| Monthly rebalance of crypto sectors based on momentum score | C | Weight distribution across multiple assets on a schedule |
| Alert when BTC drops 5% in 1 hour | B | Not a signal strategy, not a portfolio — just an alert |
| Place a limit buy order at a specific price | B | One-off execution, no signal, no schedule |
The agent will ask these questions. Having answers ready makes the conversation much faster.
What asset? BTCUSDT, ETHUSDT, a Taiwan stock code, WTI crude (CL)? The asset determines what data is available and which exchange API the agent will use.
What time interval? 5min, 1h, 4h, 1d? Shorter intervals mean more trades and fees; longer intervals mean slower reaction but less noise. For a first strategy, 1h is a good balance.
What signal idea? You don't need to know the code, just the concept: "moving average crossover", "go long when RSI is oversold", "follow the Blave Taker Intensity indicator". The agent will handle the implementation.
Spot or futures/perpetual? Futures allow leverage and shorting; spot does not. The agent will ask this before deploying live, but knowing early helps it write better code from the start.
If you're genuinely unsure where to begin, the simplest meaningful Type A strategy is an SMA crossover on BTCUSDT 1h:
It won't make you rich, but it will teach you the entire workflow — backtest, parameter scan, heatmap, deployment confirmation — before you invest time in a more complex idea.
← Back to Learn
