Reading a copy-trading backtest means knowing which numbers actually decide whether a strategy was good — and which ones are just there to flatter you. A green finish is the easiest result to misread. The wallet you tested might have beaten a coin flip and still lost to simply holding SOL; it might have ended up while dipping so hard mid-window that you'd have closed it in a panic. The output tells you what would have happened. Reading it well is a separate skill.
This guide walks each part of a Soltrace backtest and what "good" looks like for it — the return and its baseline, the win rate, the drawdown, realized versus unrealized P&L, the funnel, the per-wallet split, and the fee-and-rug line — then how to spot a result that's fragile or over-fit. If you haven't run one yet, start with how to backtest a Solana copy-trading strategy and come back with a result in front of you.
Start with return — but against a baseline
The first number everyone looks at is the total return, and on its own it's close to meaningless. Up 30% sounds great until you ask: up 30% compared to what?
The honest comparison is buy-and-hold. If SOL — or the tokens the wallet was trading — rose 40% over the same window, a strategy that returned 30% didn't add anything; it subtracted, at more risk and more effort. Beating the market is the bar, not being green.
- Green isn't the same as good. In a rising market almost everything is green. The question is whether the strategy outperformed the boring alternative of doing nothing.
- Red isn't automatically bad. A strategy that lost 5% while the tokens it touched fell 40% was doing real defensive work. That's a wallet-and-rules combination worth a second look.
- Read return next to drawdown, always. A finish means nothing without the path it took to get there — which is the next number.
For the wider question of whether any of this clears the bar over time, see is Solana copy trading profitable.
Win rate versus average return
Win rate — the share of copies that closed in profit — is intuitive and routinely misleading. A low win rate can still be a winning strategy, and a high one can quietly bleed you.
What matters is win rate paired with average return. A wallet that wins 30% of the time but whose winners are large and losers are capped can crush a wallet that wins 70% of the time with tiny gains and the occasional catastrophic loss. On Solana this pattern is everywhere: launchpad-style strategies miss often and pay for it with the rare 10x.
- Low win rate, big average winner. Fragile-looking, often durable — as long as the winners aren't a single trade (more below).
- High win rate, small average winner. Comfortable to watch, dangerous if one loss can erase a month of small gains.
- Read the distribution, not the headline. Soltrace lists every fill and exit with its reason, so you can see whether profit came from many trades or leaned on a handful.
Max drawdown is the survival number
If you read one number honestly, make it max drawdown — the deepest peak-to-trough fall along the equity curve. It's not a measure of return; it's a measure of whether you'd have stayed in the seat.
A strategy that finishes up 50% but drops 60% somewhere in the middle is, for most people, a losing strategy — because almost nobody holds through a 60% drawdown without closing the vault and walking away. The finish you never reached doesn't count. Drawdown is the number that tells you which results you could actually have lived through, and it should carry as much weight as the return, not less. The same discipline shows up in position sizing: the sizing rule that halves your worst dip is usually the one you'd have kept.
Realized versus unrealized P&L
A backtest window ends on a specific second, and some positions are still open when it does. Those open positions get marked at their last price — which means part of your result may be paper gains that were never actually banked.
This flatters results in a predictable way. A window that closes at a local top shows fat unrealized profit that a single next-day dump would have erased. So separate the two:
- Realized P&L is what your exit rules actually closed. It's the part that survived contact with your take-profits, stops and trailing stops.
- Unrealized P&L is open positions marked to the last price. Treat it as provisional, not banked.
- Discount a result that leans on open positions. If most of the gain is unrealized, you tested a strategy that got lucky with when the window ended, not one that closed well.
The funnel — why most buys got skipped
Soltrace shows a funnel: every buy it observed from the target, how many passed your filters, and how many were actually copied — with a reason attached to each skip. This is one of the most useful and least-read parts of the output.
Most observed buys should get skipped. Your filters — minimum liquidity, minimum token age, authorities revoked, max price impact — exist precisely to drop the copies you'd never have wanted. Seeing the funnel narrow is the filters doing their job. But read the shape:
- A wide skip rate with a clean result means your filters cut risk without cutting much return. That's exactly what you want.
- Copying almost nothing is its own warning. A strategy that passed three trades out of two hundred isn't robust — it's a tiny sample that got lucky, and it'll behave unpredictably live.
- Read the skip reasons. If "min liquidity" is dropping everything, you may be too strict for this wallet; if "authorities revoked" is catching a lot, that's a rug filter earning its place. This is the same funnel you compare filters-on against filters-off in the building blocks of a strategy.
The per-wallet breakdown
If you tested more than one wallet, the aggregate result hides who did the work. The per-wallet breakdown splits the outcome out, and it routinely tells a different story than the headline.
One wallet can carry an entire multi-wallet test — or sink it. A blended "up 20%" might be four flat wallets and one monster, which means you don't have a five-wallet strategy, you have a one-wallet strategy wearing a disguise. Drop the carrier and the result collapses.
- Find the carrier. If a single wallet produced most of the gain, your real strategy is that wallet — size and judge it accordingly.
- Find the anchor. A single wallet dragging the group down is often better removed than tolerated for diversification's sake.
- Don't average away the truth. Per-wallet is where you see whether "copy ten wallets" was actually diversified or just one bet with nine spectators.
Fees and rug count
Two lines close out an honest read. The fee total — Soltrace charges a flat 0.85% per trade, no profit share — should already be baked into the equity curve, but check that a high-frequency strategy isn't handing back its edge in per-trade cost. A wallet that trades constantly needs a bigger gross edge to clear the same net.
The rug count is the tally of copied tokens that went to near-zero. A low count next to your filter settings tells you the safety rails held; a high one says your entry filters were too loose for what this wallet trades, and it's worth tightening them and re-running.
Spotting a fragile or over-fit result
The dangerous backtests aren't the ugly ones — they're the pretty ones that won't survive contact with next week. Three tells:
- One-trade dependence. Strip out the single best trade. If the strategy goes from great to mediocre, you didn't find an edge; you found a lucky entry, and it won't repeat on schedule.
- A great finish hiding a brutal drawdown. Covered above, and it bears repeating because it's the most common self-deception: the curve you'd never have held through isn't a result you can bank on.
- Works on exactly one week. A strategy that only shines on your chosen window is probably fitted to that window's noise. Re-run across a different — ideally wider — window and see if it holds. How far back to backtest covers picking windows that expose this, and can you backtest a Solana wallet covers what's testable in the first place.
From result to live
Once you can read the whole output — return against its baseline, win rate against average size, drawdown as the survival test, realized separated from paper, the funnel and per-wallet split, fees and rugs — you know whether a strategy earned live funds or just flattered you. From there the move is short: assemble the rules in the strategy designer, fund a non-custodial Solana PDA vault you own the whole time, and let the keeper run the same rules forward. The keeper can place and size copies but can never withdraw; execution is keeper-signed across the Jito, Nozomi, Astralane and AllenHark relays, with optimistic execution on Pro and Elite, at a flat 0.85% per trade and no profit share.
Read the result honestly and it does one job well: it stops you funding a strategy you couldn't actually have held. That's worth more than any single green number.
Ready to read a real one? Run a Soltrace backtest and judge a wallet on the full picture before you fund a thing.
Copy trading does not guarantee profit, and crypto trading carries substantial risk of loss. A backtest describes the past under your rules; it is not a prediction. Nothing here is financial advice.