Core Mental Model
The bot is not trying to predict every tiny move. It is managing inventory under uncertainty.
Think of every position as a small business decision:
Buy only when there is a reason.
Hold only while the reason remains true.
Sell when the reason is false, the risk is too high, or the planned profit/risk logic says to exit.The Three Questions
Every trade should answer three questions before entry:
| Question | Meaning | Bad answer |
|---|---|---|
| Why buy? | The setup or thesis | “Price is down” |
| Why now? | Timing and confirmation | “It might go up” |
| Why exit? | Invalidation and profit plan | “When green” |
Example:
Asset: BTCUSDT
Thesis: Price is near support, 1h trend is stable, RSI suggests oversold conditions.
Entry: Small long only if BTC holds support for another cycle.
Invalidation: Close/reduce if price breaks support or RSI weakness continues.
Profit plan: Take partial profit near prior resistance; trail the rest if trend strengthens.
Timeout: Exit or reassess if nothing happens after 6 cycles.Spot Trading Is Inventory
On spot, the bot owns coins. It can:
- Buy more.
- Sell some.
- Sell all.
- Hold cash.
- Rotate between assets.
It cannot open a true short. If the bot thinks ETH will fall, the spot action is not “short ETH.” The action is reduce ETH, close ETH, hold USDT, or wait for a better long setup later.
Movement Is Not Edge
Seeing that price is up or down is not enough. Everyone sees that.
The bot needs an edge like:
- Price is at a level where reward is meaningfully larger than risk.
- Multiple timeframes agree.
- The market regime supports this setup.
- The portfolio has room for the trade.
- The expected move is larger than fees, spread, and slippage.
The Best Default Is Patience
A 24/7 bot does not need to trade constantly. It has the luxury of waiting.
The right identity is:
I can act any time, so I do not need to act every time.Gotcha: “No trade” is not failure. A bad trade avoided is a win.