Turn your trading approach into an AI agent. Build · Test · Deploy. Crypto. Early access
New York, USA
Joined February 2013
- Tweets57
- Following21
- Followers10
- Likes26
Cook 100 agent swarm market research takeaway today: DO NOT SELL.
65% likelihood another short term leg down in AI and crypto markets.
Scale in long at the next resistance.
Everyone's asking if the AI bubble popped.
"Bubble" was called in 1996.
The Nasdaq peaked in 2000.
You learn a bubble burst the way you learn you overslept.
Don't time the pop. Size for it.
What's your exit rule if it pops tomorrow?
What to watch instead of headlines: leverage piling up and everyone crowded into the same trade. Cook turns your rules into an agent that alerts you, and only acts inside limits you approve. Early access → cooktrades.com/access?utm_so…
A backtest without fees is a fantasy.
Make 0.1% per trade, pay 0.1% in fees and slippage, and you've built a hobby, not an edge.
Do your backtests include fees?
In 2012 a cat picked stocks by batting a toy mouse onto a grid.
It beat a team of pros: +10.9% vs +3.5%.
1,000 people flipping 10 coins each: 62% chance one gets 10 heads in a row.
How many trades before you'd trust a streak?
Luck and skill look the same over a short record. Cook backtests your rule across hundreds of trades, fees and modeled slippage in. Evidence, not proof. Early access → cooktrades.com/access?utm_so…
In Cook, approving a strategy and allowing live orders are two separate decisions.
You can love the rules and still keep the agent on alerts only.
Would you start with alerts or live orders?
17 indicators.
0 rules.
RSI says buy. MACD says sell. The cloud says maybe. 🎄
Indicators are ingredients. Rules are the recipe.
How many indicators are on your chart right now?
Cook's 3-line version: trend (price above the 50-period MA), trigger (a 4h close above resistance), risk (1% per trade). Simple enough to backtest. Build yours → cooktrades.com/access?utm_so…
Supertrend flipped red. Sell everything?
It's ATR × 3 trailing price. It lags, and in sideways chop it flips again and again: a small loss each time.
A seatbelt, not a steering wheel.
What do you use to filter the chop?
Cook's way: write your Supertrend rule plus a sideways filter in plain words, then backtest it with fees before trusting it. Early access → cooktrades.com/access?utm_so…
RSI 70 isn't a sell signal.
It only says recent moves were strong. In a strong trend, RSI can stay above 70 for weeks.
A spice, not the recipe.
What do you pair RSI with?
Cook's way: put RSI inside a rule with something it can't see on its own (trend, a level), then backtest the whole recipe with fees before trusting it. Early access → cooktrades.com/access?utm_so…
Risk 1% doesn't mean invest 1%.
It means: if your stop hits, you lose 1% of the account. The stop distance sets the size.
Do you size from the stop, or from a feeling?
This trading robot lost $440 million in 45 minutes.
Aug 1, 2012: Knight Capital's new code reached 7 of 8 servers. The 8th ran code from 2003.
It wasn't evil. It had no limits.
What's the first limit you'd give a trading bot?
About 4 million trades in 154 stocks before anyone stopped it. Cook runs agents only inside limits you set, and live orders need your OK. Early access → cooktrades.com/access?utm_so…
Checking the chart 47 times doesn't make the setup arrive faster.
It just makes you take the ones that aren't there.
Alerts > attention.
A 50% loss needs a 100% gain to get back to even.
A 20% loss needs 25%.
That gap is why sizing matters more than entries.
What's your max loss per trade?
47 chart checks today.
BTC moved 0.3%.
Send this to the friend who "just checks the chart" 📱
How many times did you check today? Be honest.
The fix: put an alert where your setup actually is (say, a 4h close above your level), then close the app. Cook agents watch your watchlist and ping you when your rules say it's ready. Early access → cooktrades.com/access?utm_so…