The losses nobody talks about
Scams get headlines and crashes get think-pieces. Most money leaks out through defaults nobody revisited.
What are the hidden costs of crypto trading?
The main hidden costs are choosing the default transfer network, sending during network congestion, instant-convert spreads, sandwich attacks on decentralised swaps, paying taker fees where limit orders would do, and funding on positions held too long. Together these often exceed a trader's actual losses.
Ask what causes crypto losses and you will hear about hacks and crashes. The research says otherwise: most losses come from operational mistakes — wrong networks, hidden fees, bad withdrawal timing, spreads and address errors.
Each of these is a decision made by default rather than on purpose. The default was set once for a reason and never revisited — and in crypto, an unconsidered default is paid immediately and in full.
1. The default network
Whatever the exchange preselects is not necessarily the cheapest or the one your receiver supports. Moving USDT on Tron costs about a dollar; the same transfer on Ethereum can cost twenty. Over a year of regular transfers, the default alone is a meaningful sum. See the network guard before every send.
2. Withdrawal timing
Bitcoin network fees are set by an auction for block space, and that auction has quiet hours. Sending during congestion can cost several times what the same transaction costs a few hours later. Unless the transfer is urgent, the send timing panel will tell you whether waiting is worth it.
3. The conversion spread
Instant-convert buttons and card purchases quote a price that already includes a margin. It is not called a fee, so it does not appear in any fee schedule, and it is frequently many times larger than the trading fee you would pay placing the same order on the exchange's normal market.
4. Sandwich attacks on decentralised exchanges
5. Copying from history
Reusing an old address from your transaction history feels safe and is one of the ways clipboard-hijacking malware succeeds. Addresses should be verified at the moment of sending, every time, first and last characters checked visually.
6. Taxes discovered too late
In many jurisdictions, trading one crypto for another is a taxable event even though no money left the system. The tax problem is fundamentally a data problem: the right moment to fix it is now, while you can import history cleanly and your memory of the transactions is fresh, rather than under deadline pressure months later. The tax export panel produces a clean spreadsheet from your journal and holdings in one click.
The common thread
Every item on this list shares one shape: a decision made by default rather than by intention. Traditional finance has correction mechanisms — reversals, dispute windows, support teams who can pull funds back. Crypto has none of them. That is precisely why the five minutes spent checking beforehand is worth more here than anywhere else.
Putting a number on each leak
Annual cost for an active retail account
| Default network on transfers | $50–300 |
| Sending during congestion | $30–200 |
| Instant-convert spreads | $100–1,000+ |
| Sandwich attacks on DEX swaps | 0.3–2% of every swap |
| Taker fees where limits would do | $150–600 |
| Funding on positions held too long | often the largest of all |
Individually each looks negligible, which is exactly why none of them is examined. Together they routinely exceed what the same trader loses on bad trades — and unlike bad trades, every one is avoidable with a habit rather than a skill.
The instant-convert spread, examined
Exchange "convert" buttons and card purchases quote a single price with the margin already inside it. Nothing is labelled a fee, so nothing appears in any fee schedule — and the margin is frequently several times the trading fee you would pay placing the same order on the normal market.
The test takes ten seconds: note the price the convert button offers, then look at the mid-price on the spot market for the same pair. The difference is what the convenience cost. For most people it is between 0.5% and 2%, on every single conversion.
Sandwich attacks, in detail
On a public network your pending swap is visible before it executes. A bot sees a large swap heading for a pool, buys immediately before it to push the price up, lets your order fill at the worse price, and sells straight after — capturing the difference.
You cannot prevent it, but three things limit it substantially:
- Set slippage tolerance tight. Your tolerance is the ceiling on how much can be extracted. A default of 5% is an invitation.
- Split large swaps. Several smaller trades are less attractive targets than one large one.
- Prefer deeper pools. The attack works by moving price; deep pools move less.
Withdrawal timing, quantified
Bitcoin fees are an auction for limited block space, and that auction has predictable quiet periods — typically weekends and the hours around 02:00 to 08:00 UTC.
The gap between the fast tier and the economy tier is frequently three to five times. On a transfer that is not urgent, checking the send timing panel and waiting a few hours is among the highest hourly rates you will ever earn for doing nothing.
The same logic applies to Ethereum and its layer-two networks, where the gas panel shows six chains at once. The difference between the most and least expensive chain for the same swap is often a hundredfold.
The tax problem, and why it is really a data problem
In most jurisdictions, trading one crypto asset for another is a disposal and therefore a taxable event, even though no fiat currency moved. Airdrops are commonly taxable as income at their value on receipt, and selling them later creates a second, separate gain.
None of that is difficult while the records exist. It becomes very difficult months later, when exchange history has been partially deleted, a venue has closed, and you cannot remember what a transfer was for.
Export quarterly rather than annually. The tax export panel builds a clean spreadsheet from your journal and holdings inside your browser, and the right time to run it is while the trades are still fresh.
A quarterly audit worth doing
- Add up every fee and funding payment for the period. Compare it with your net result.
- Check which network you have been defaulting to for transfers.
- Review outstanding token approvals and revoke unused ones.
- Compare your average fill price against the price at the moment you clicked, on ten trades.
- Export your records for tax while the memory is intact.
An hour, four times a year. For most active traders it recovers more money than any improvement to their entries.
Common questions
Are these costs unique to crypto?
The categories exist everywhere; the difference is that crypto has no correction mechanism. Traditional finance has reversals, dispute windows and support teams who can pull funds back. Here, a mistake is final, which raises the value of checking beforehand enormously.
Which single change saves the most money?
For active traders, checking funding before multi-day holds. For everyone else, using the spot market instead of the convert button. Both take seconds and both are usually worth more than any improvement in trade selection.
Do I need tax software?
Not necessarily, but you do need records. If you trade frequently, software that imports exchange history saves considerable time. If you trade rarely, a quarterly spreadsheet export is sufficient — the failure mode is having no record at all.
Free, no account, nothing stored on our servers.
Open Preflight