DEV Community

Gennady
Gennady

Posted on • Originally published at trendrider.net

How I Pick Crypto Trading Pairs for My Bot — A Data-Driven Framework

Choosing the right trading pairs is one of the most underrated aspects of building a profitable crypto bot. After months of testing, here's the framework I use to select and rotate pairs.

Why Pair Selection Matters

Most tutorials focus on entry signals and indicators. But trading the wrong pairs can kill even a great strategy. A perfect RSI reversal signal on a low-liquidity altcoin will get eaten by spread and slippage.

My bot trades 15 pairs on Bybit futures. Here's how I picked them.

The 5 Criteria I Use

1. Minimum Daily Volume: $50M+

Anything below $50M in 24h volume means:

  • Wide spreads that eat your profits
  • Slippage on entries and exits
  • Gaps that trigger false signals

I check volume on CoinGecko and cross-reference with Bybit's actual order book depth.

2. Volatility Sweet Spot: 2-5% ATR

Too low volatility (BTC in a tight range) = no opportunities. Too high (meme coins doing 50% swings) = stop losses get destroyed.

I measure ATR as a percentage of price on the 1h timeframe:

atr_percent = (ta.ATR(dataframe, timeperiod=14) / dataframe['close']) * 100

# Sweet spot: 2-5%
if 2.0 <= atr_percent <= 5.0:
    pair_score += 2
Enter fullscreen mode Exit fullscreen mode

Most major altcoins (SOL, ETH, BNB) sit in this range. DOGE and meme coins often exceed it.

3. Spread < 0.05%

Spread is the hidden fee. If your average trade makes 1% and the spread is 0.1%, you're giving away 10% of your edge on every round trip.

I only trade pairs where the typical bid-ask spread stays under 0.05%. The top 15 by market cap almost always qualify.

4. Low Correlation With Each Other

Trading BTC, ETH, SOL, and BNB might feel diversified, but when BTC drops 5%, they ALL drop. I measure 30-day rolling correlation:

Pair BTC Correlation
ETH 0.85
SOL 0.78
BNB 0.72
DOGE 0.65
LINK 0.61
ATOM 0.55

Lower correlation = better diversification. I include a mix: some high-corr (ETH, SOL) for trend-following, some low-corr (ATOM, NEAR) for mean-reversion.

5. Backtest Validation

Every pair must pass backtesting before going live:

  • Minimum 50 trades over 3 months
  • Win rate > 55%
  • Profit factor > 1.5
  • Max drawdown < 3%

If a pair fails any criterion, it doesn't make the cut. Period.

My Current 15 Pairs (Tiered)

Tier 1 — Core (highest confidence):
BTC, ETH, SOL, BNB

Tier 2 — High-value altcoins:
DOGE, XRP, ADA, AVAX, LINK

Tier 3 — Diversification:
DOT, POL, NEAR, ATOM, SUI, OP

Pairs I Avoid

  • New listings (< 3 months): Not enough data to backtest
  • Meme coins (except DOGE): Unpredictable pump/dump patterns
  • Low market cap (< $500M): Liquidity disappears during volatility
  • Delisted/rebranded: Recently had to swap MATIC→POL and FTM→SUI

Monthly Rotation

Every month I:

  1. Re-run backtests for all 15 pairs
  2. Check if any pair's volume dropped below threshold
  3. Evaluate 2-3 new candidates
  4. Replace underperformers

This keeps the portfolio fresh without constant tinkering.


Results

With this framework:

  • 67.9% win rate across all 15 pairs
  • 2.12 profit factor
  • 1.42% max drawdown

The pair selection contributes as much to these numbers as the actual trading signals.


What pairs does your bot trade? How do you select them?

I share all trades publicly: @TrendRiderFree on Telegram

Top comments (0)