Practice profits are not real profits: why we keep testnet and live separate

14 September 2026 6 min read Stralines

#education
#testnet
#risk
#honesty

You ran a bot on a practice account for three weeks. The ledger is green. It is very hard, at that point, not to feel like a trader who is up.

That feeling is the problem this post is about. Practice mode exists so you can pressure-test a strategy with zero capital at risk, which is genuinely valuable. But a green practice week is not earnings, and the moment those numbers start blending into your sense of how I’m doing, you are measuring an account that does not exist.

This is educational only. It is about how to read your own records honestly, not about what to trade.

The trap is psychological, not technical

Nobody sets out to confuse the two. It happens because the number looks identical in both places — same currency, same formatting, same little green arrow. Nothing in the interface tells you that one of them is rehearsal.

Four things make practice results systematically kinder than live ones:

Fills are too generous. A practice order does not compete for real liquidity. On many testnet venues the book is thin or synthetic, and you get filled at prices a real book would not have offered you — especially at size, and especially in the fast moves where your strategy makes or loses most of its money.

Slippage is understated. Related but distinct: even where a practice venue models slippage, it models an average rather than the specific awkward moment your strategy tends to trade in. Systems that trade breakouts are exactly the systems that meet the worst spreads.

Costs may not match. Funding rates, maker/taker tiers and fee schedules are not always mirrored faithfully on a practice venue. Small per-trade differences compound quickly across a few hundred trades.

You were not afraid. This is the biggest one and it gets discussed least. On a practice account you did not hesitate before the entry, did not close early because the drawdown felt unbearable, did not quietly reduce size after three losses. The strategy was tested. You were not. And in live trading, the operator is part of the system.

Practice mode measures the strategy’s mechanics. Live trading measures the strategy plus the venue plus you. Those are different distributions, and the gap between them is not a rounding error.

Two ledgers that must never be added together

If you add a practice P&L to a live P&L, the sum describes no account anywhere in the world. It is not an estimate, an approximation, or a best-case — it is a number about nothing.

So the two have to stay apart, and the cleanest place to enforce that is at the source rather than in the reporting layer. On Stralines every signal carries its own record — its own position, its own orders, its own realised P&L — so practice and live activity are never pooled into a shared figure that then has to be carefully unpooled later. There is no combined total to accidentally surface, because one was never computed.

That is a deliberately boring design choice, and it is load-bearing. A system that mixes the two at the source and separates them at display time will eventually leak, because every new report is another chance to forget which is which.

What this looks like in the product

Three places where the separation is visible rather than implied:

Performance emails headline the live number only. Practice-account activity appears in a clearly labelled block that is excluded from any “here is what your account did” figure. It is reported, not counted.
Plan tiers cap practice and live bots separately. Every tier lists its practice allowance and its live allowance as distinct limits, so moving a strategy from one to the other is a decision you make, not a side effect of deploying one more bot.
The free tier leans on practice. Explorer includes five practice bots and a single live Futures bot. You can run the whole system, in earnest, for as long as you want, with at most one strategy’s worth of market exposure — and without a report that flatters you into thinking practice time was profit.

Why we would rather show a smaller number

There is an easier version of this product. Add both ledgers together, put one large total at the top of the email, and let it feel good. Retention would probably improve for a quarter.

It would also mean the number we send you is the number least connected to your actual position — and the person most misled by it is the person paying us. A platform that trains you to trust an inflated figure has made you worse at trading while charging you for the privilege.

So the headline stays live-only, even when live-only is less impressive than the combined figure would have been. Honest reporting is a product feature here, not a compliance tax we grudgingly pay.

How to use practice mode well

Practice is a rehearsal for process, not a preview of return. A checklist that makes it genuinely useful:

Confirm the risk caps actually bite. Set a position cap and a drawdown limit deliberately tight, then verify the system refuses to exceed them. You want to see the guardrail work before you need it.
Confirm protective orders place and hold. Check that stops and targets land on the venue and stay there through position changes — the failure mode covered in why your stop-loss disappears.
Sit through a losing streak without touching anything. This is the real test, and the only part of practice that rehearses you. If you override the system on practice money, you will certainly override it on real money.
Learn the shape, not the total. Note the maximum drawdown, the longest losing run, and the worst week. Those transfer to live trading far better than the P&L does. The 35% win rate post is about reading exactly these numbers.
Size down for your first live weeks, and start the ledger at zero. Not just in the platform — in your head. Judge live on its own record from day one, with no credit carried forward.

The goal of practice is a calmer live account. It is not a bigger imagined balance.

Closing

Practice mode is one of the few tools in trading that lets you be wrong for free. That value is entirely destroyed the moment you start counting the results as money, because then you are not learning from the rehearsal — you are drawing conclusions from a performance nobody paid for.

Keep the two ledgers apart. Judge the practice run on whether your process held, and judge live on live.

Stralines is a software automation platform. We do not provide investment advice, manage funds, or execute trades on our own books. All trading occurs in your own exchange account using your own API keys. Past performance shown in backtests and simulations does not predict future results. Trading cryptocurrencies involves substantial risk of loss.