hummingbot

Běžel jsem market making bot na Binance s malým účtem (~$600) a tady je upřímná pravda 👇
📊 Co se vlastně stalo:
Několik seancí
Nějaké BUY/SELL cykly
Čistý PnL ≈ na nule až malý zisk
Ale tady je ta zajímavá část…

📉 SOL spadl o ~2%
🤖 Můj bot: téměř plochý
💡 Takže jo… NEztratit peníze v klesajícím trendu už vypadá jako výhra.
⚠️ Největší chyby, které jsem udělal (a vidím, že je mnoho lidí opakuje):
❌ Spread příliš široký → 0 vyplnění → “mrtvý bot”

❌ Nakupovací skew blokující BUY → jen prodeje

Yesterday I shared my first results.
Today — things got real 👇
📊 What changed:
❌ Before:
• 0 trades
• Wide spreads → no execution
• Bot looked “smart”… but did nothing
✅ Now:
• 20+ trades in session
• Balanced BUY/SELL cycles
• Real market interaction
But here’s the twist…
💰 PnL Result:
• Trade PnL: positive
• Total PnL: slightly negative
🤔 Why?
Because of one brutal truth:
⚠️ Fees > Edge
Even with perfect execution, small spreads (0.1–0.2%) get eaten by fees.

🔥 Biggest realizations so far:
1️⃣ A bot that doesn’t trade = useless
2️⃣ A bot that trades too cheap = also useless
3️⃣ Execution is step 1… but EDGE is step 2

📉 Market context:
• SOL moving in tight range (~82–84)
• Low volatility → harder to capture profit
• High competition → spreads must be precise

🧠 What I changed:
✔ Removed artificial spread distortion
✔ Fixed inventory skew behavior
✔ Increased effective spread (0.25%+)
✔ Focused on PnL per TIME, not per trade

📊 Current goal:
👉 Fewer trades
👉 Higher profit per cycle
👉 Positive net PnL after fees

💡 My current hypothesis:
“More fills ≠ more profit
Better edge + controlled fills = profit”

🤔 Question for traders & bot runners:
What matters more for you?
1️⃣ High frequency (many trades, small edge)
2️⃣ Lower frequency (fewer trades, higher edge)

📌 Next step:
Testing dynamic spreads based on volatility (ATR-based)
Will update results soon 👇
#Hummingbot #MarketMaking #SOL #CryptoTrading #AlgoTrading