How to Copy Trade on Solana: Complete 2026 Guide

1. Pick the right wallets (this is 80% of the game)

Copy trading is a wallet-selection problem. Your execution bot is the easy part — picking wallets that keep printing is what actually matters.

Public Solana leaderboards like kolscan.io, Dexscreener's smart-money feeds, and Subglow's own /trader leaderboard surface the top wallets by realized PnL over 24h / 7d / 30d windows. That's your starting pool. But raw PnL alone is misleading — a wallet can land one 100x and sit in the #1 slot for a week while the rest of their trades print red.

• Appears on 7-day AND 30-day leaderboards, not just daily — signals consistency over luck.
• Win rate above 40% (memecoin trading is asymmetric — even 30% hit rate with good sizing works).
• Trade count above ~20 over the period you're evaluating — fewer than that is too small a sample.
• Doesn't round-trip entire positions — check their transaction history on Solscan. Multi-tranche exits beat all-or-nothing exits for risk-adjusted return.
• Has a clear visible style (e.g. only fresh Pump.fun launches, or only established $1M+ market cap tokens). Style drift = declining edge.

2. Set position sizing before you enable copying

The most common rookie mistake: copying a whale at 100% of whatever they trade. If they put 50 SOL into a coin and you have a 10 SOL total portfolio, you're instantly margin-called on a single bad call.

Subglow supports three sizing modes:

1. Fixed amount. Set a constant SOL amount per copied trade (e.g. 0.2 SOL). Simple, predictable, and what we recommend for beginners. Use this for your first 2 weeks while you learn the wallet's cadence.
2. Mirror percent. You copy a fixed percentage of the whale's trade size (e.g. 5%). Good once you trust the wallet — your position scales with theirs.
3. Scaled tiers. Different sizes per whale trade-size band (e.g. '0.1 SOL if whale <1 SOL, 0.5 SOL if whale >5 SOL'). Useful for whales who mix exploratory small bets with conviction large bets — you only want to copy the conviction bets.

3. Turn on the filters that save you from rugs

On any given week, 30-60% of the new memecoins a whale buys are rugs-in-waiting. The whale doesn't care — they exit in 3 minutes. You, copying 30 seconds later with a $0.5 slippage, absolutely do.

Three filters cut rug exposure by 80%+:

• Anti-rug level: moderate or strict. Blocks tokens with mutable mint/freeze authority, locked LP below a threshold, or suspicious holder concentration.
• Cooldown: 300–600 seconds. Prevents the bot from re-buying the same token after the whale dumps and re-enters (common rug-pump-dump pattern).
• Min trade size: 0.1 SOL. Filters out whale dust/probing trades that are usually noise.

4. Set TP / SL independently of the whale

Don't rely on the whale to exit for you. Whales hold longer than retail copy-traders should, because they're diversified across 50+ positions and can stomach volatility. You probably can't.

A reasonable starting set of exits for memecoin copy trades:

• Take-profit (TP) at +50% to +100%. Close the whole position. Boring, but 90% of copy-traders never take profit and watch 2x's round-trip to 0.
• Stop-loss (SL) at -30% to -40%. Hard floor. A tight SL gets whipsawed; a loose SL turns into a bag. 30-40% is the sweet spot for Solana memes.
• Trailing stop: 20%. Ratchets up as price rises. Locks in profit on runners without capping upside.

5. Portfolio-level risk caps

Per-trade sizing is necessary but not sufficient. You also need caps on total exposure — otherwise a bot copying 10 wallets will concentrate all 10 into the same token the moment they all FOMO in.

Subglow's portfolio caps:

• Max open positions: hard cap on total number of open positions. 5-10 is a reasonable range.
• Max per-token SOL: hard cap on exposure to any single token, regardless of how many wallets bought it.
• Portfolio % limit: no single token can exceed X% of your total copy-trade portfolio. 10-15% is a common setting.

6. Run it for a week before scaling

New copy traders make one of two mistakes: scale up after 2 good days, or turn it off after 2 bad days. Both are reactions, not decisions.

Run for at least 5-7 days with consistent settings before adjusting. That gives you enough trades to see the wallet's real hit rate, how your filters interact with their style, and whether your TP/SL thresholds fit their cadence.

What changes in 2026 vs. older copy-trade playbooks

• Execution latency matters more. gRPC-based execution (sub-second detection + Jito bundles) has replaced webhook-based bots. The 500ms difference between systems now routinely costs you 10-20% slippage on fresh launches.
• Proportional sells are now table-stakes. Any bot still doing 'sell 100% when whale sells' is behind.
• Smart-money confluence is the new alpha. Copying any one wallet is 2024. Copying 'when 2+ tracked wallets buy the same token within 2 minutes' is where edge has migrated.
• Flat fees are in, per-trade + priority-fee markups are out. The market has repriced — 0.9% per trade doesn't fly anymore.