🔎How to verify the returns a platform reports: six checks, and the one you cannot make

The annualised number on your dashboard was calculated by the platform. That is true of any custodial platform, ours included.

So the useful question is not “can it be trusted” but: which numbers can you check yourself, and which can you not. This piece draws that line.

The six things you can pull

These six are available to an investor in real time:

Available	What it answers
Balance	What is in the account right now
Trade-by-trade detail	What each individual movement was
Positions	What each leg is holding at this moment
Return history	The return curve, period by period and stretch by stretch
Subscription records	When you entered, and on which plan
Withdrawal progress	Whether the money has left, and how far along it is

On the backend these sit behind one authorisation group that the investor themselves can call. They support more reconciliation work than they look like they do — the rest of this piece is how to use them.

Step one: recompute the funding rate from the exchange’s public API

This is the hardest link in the chain, because it does not depend on the platform giving you anything.

Funding rates are public. Anyone can pull them. On Binance USD-M the per-period history is at fapi/v1/fundingRate — no key, no login.

Here is the pull we did ourselves (1 January to 21 September 2026 00:00 UTC, 790 periods per instrument):

Instrument	Year to date	Annualised	Negative periods
BTCUSDT	2.103149%	2.92%	209
ETHUSDT	1.277538%	1.77%	247
SOLUSDT	−0.941791%	−1.31%	390
BNBUSDT	2.078423%	2.88%	7

Hold that table against whatever a platform reports, and the point is not matching to the decimal, it is whether the order of magnitude is reachable. Funding is the only revenue source this approach has; the market pays what it pays, and nobody conjures a figure an order of magnitude above it.

When you see a number clearly out of line with the prevailing market level, you do not need to judge it true or false first. You just need one explanation: where does the extra come from. Legitimate answers exist — different instruments, different notional, other revenue lines folded in — but somebody has to be able to state it.

How the rate itself works and why it was so low this year is in 790 periods, not one above 0.01%.

Step two: back out the notional from trade-by-trade detail

Return = rate × notional. You already have the rate independently, so the notional falls out:

implied notional = funding credited that period ÷ that period's public rate

Compare the implied notional against the notional shown on the positions page. Those two should agree. The three usual reasons they do not are all answerable: the position was adjusted mid-period, it is split across sub-accounts, or the period included another instrument.

The value of this step is that it breaks the aggregate called “return” back into two factors you can check separately. Aggregates are hard to verify; factors are easy.

Step three: run the plan arithmetic yourself

Between gross return and what lands in your account sits a fixed set of arithmetic, and it should be public and reproducible. Ours reads:

management fee   = principal × 2%           ← same on all three plans, taken from returns, never billed to you
distributable    = gross − management fee − floor (zero if it does not cover)
you receive      = floor + distributable × your share

Three numbers define the three plans:

	Principal guarantee	Floor (net to you)	Excess split (platform : you)
Plan 1, principal + return	Yes	7%	100 : 0
Plan 2, principal, 7:3	Yes	None	70 : 30
Plan 3, no principal guarantee, 3:7	No	None	30 : 70

Run 100,000 USDT through it at a 10% gross strategy return:

	Gross	Fee	Floor	Distributable	Your share	Net to you
Plan 1	10,000	2,000	7,000	1,000	0	7,000 (7.0%)
Plan 2	10,000	2,000	0	8,000	2,400	2,400 (2.4%)
Plan 3	10,000	2,000	0	8,000	5,600	5,600 (5.6%)

Working it through surfaces something rarely spelled out: which plan wins depends entirely on where gross return lands, and the crossover points can be solved for.

• At 12% gross, plan 3’s net exactly catches plan 1’s 7%; below 12%, the floor plan pays more.
• Plan 2 needs 25.33% gross before it matches that same 7%.

Neither number is a marketing choice; both drop out of the three lines of arithmetic above. You should be able to solve for the same two numbers yourself — and if a platform’s split arithmetic cannot be solved, that is not your failing.

Two things that are easy to misread, while we are here:

• Principal guaranteed is not return guaranteed. Plan 2 guarantees principal, not return: the principal comes back whole, and what it earned that year follows the strategy.
• Custody is not a principal guarantee. All three plans hold funds in a platform account, but only plans 1 and 2 carry the 100% principal guarantee.
• In a year where returns do not cover the management fee, the platform collects less and does not bill you for the difference.

Step four: test it with a withdrawal

The first three steps check numbers. Numbers can all reconcile and still not prove the money is there.

The only action that tests that is a withdrawal, and do a small one early rather than discovering it on the day you need everything out. Watch more than “did it arrive”: whether the progress states actually update step by step, and whether real elapsed time matches what the terms say.

Early-exit terms differ by plan (investment terms run from 1 to 5 years), and this is a question to settle before you enter, not on the day you want out. The full path for getting money back is in returns, early exit and withdrawals.

The one you cannot see: solvency

Do all four steps and one thing is still invisible from the investor side: the platform’s overall solvency ratio — total client liabilities against the assets backing them.

On our system that figure currently sits behind administrative authorisation. An investor cannot pull it. Saying so beats talking around it.

Why single it out: everything in your six is about your own account. All of it reconciling proves only that your slice is booked correctly. Whereas the floor and the principal guarantee are performed on the platform’s own credit, and whether it can perform depends on the whole book, not on your page.

That is the concrete shape of platform credit risk, and it does not go away. Under custody, the numbers on a dashboard are self-reported; transparency can enlarge the checkable portion, it cannot eliminate the credit risk. Our position state machine contains a real “defaulted” state, and it exists because this is not a theoretical concern.

The platform-level risks are listed under security and risk.

A checklist you can use as-is

Platform-agnostic, all eight are askable out loud:

1. Can trade-by-trade detail be exported, or only an aggregate?
2. Does a period’s credited funding, divided by that period’s public rate, back out a notional that matches the position?
3. Can the split be written as three lines of arithmetic rather than a sentence of description?
4. Is the management fee taken from returns or billed to me? What happens in a year that does not cover it?
5. Is the floor quoted net to me or on gross return?
6. Whose account holds the money? Under custody, which items can I query live?
7. What are the early-exit terms, and where are they written down?
8. Which item can I not check?

The eighth is the most useful. An answer that names something you cannot check beats an answer where everything is checkable — because the first one means they actually know where their own authorisation boundary runs.

The four questions worth settling before choosing a platform are in choose an arbitrage platform; telling real from fake is in is funding rate arbitrage a scam; our own plan terms and arithmetic are laid out on plans and fees, and the year-by-year results on our track record — including the row where ETHUSDT returned 0.787256% in 2022.

Sources: per-period funding from Binance’s public USD-M endpoint fapi/v1/fundingRate, range 2026-01-01 00:00 to 2026-09-21 00:00 UTC, 790 periods per instrument, read at 2026-09-21 07:33 UTC; plan terms (2% annual management fee taken from returns on all three plans, 7% net floor on plan 1, excess splits of 100:0 / 70:30 / 30:70, custody on all three, 5,000 USDT minimum, 1–5 year terms) per the plans and fees page; the 12% and 25.33% crossovers are solved from the arithmetic on that page and exclude compounding, since returns are settled out rather than rolled into principal; yearly totals on the track record page, on-site snapshot timestamp 2026-08-25 13:42 UTC. Past performance does not indicate future results.