APY vs total return vs drawdown: reading bot performance properly

Total return: what actually happened

Total return is the gain or loss over a period, as a percentage of what you started with:

Total return = (end value − start value) ÷ start value
Example: 10,000 USDC grows to 12,000 USDC in 9 months → +20%.

It is the most honest of the three numbers, because nothing is projected. Two things to watch:

• Deposits and withdrawals. Adding 2,000 USDC to an account is not a 20% return. Remove your own money movements before you calculate.
• The period. A total return means nothing without a start and end date.

APY: a yearly rate, not a promise

APY (annual percentage yield) restates a return as if it continued for a full year, with gains compounding. Using the example above:

APY = (1 + total return)^{12 ÷ months} − 1
(1.20)^{12 ÷ 9} − 1 = +27.5% APY

The account made 20%, but the headline says 27.5%. That is not wrong, but it assumes the next three months go as well as the first nine. APY is useful for comparing results over different periods. It is not a forecast.

Why short periods inflate APY

The shorter the period, the bigger the multiplier. A lucky 4% month becomes a 60% APY. When you see a high APY, find out how many months it is based on before anything else.

APY vs APR

APR is the simple version without compounding: 20% over 9 months is 26.7% APR. APY is always equal to or higher than APR, so check which one you are reading.

Drawdown: how bad it got along the way

A drawdown is a fall from a peak to a later low, as a percentage of the peak. The maximum drawdown is the worst one in the period.

Example: an account peaks at 12,400 USDC and later falls to 10,540 USDC. Drawdown = −15%, even if the account ends the year in profit.

Drawdown matters because losses are harder to recover than they look. After a fall, you need a bigger gain just to get back to where you were:

It is also the number that tells you whether you could live with the strategy. If you would have switched a bot off during a 25% drawdown, its long-run APY would never have reached you.

Reading the numbers together

Compare two hypothetical bots over the same two years:

Bot B made more, but at one point it had lost almost half the account. Anyone who joined near that peak waited a long time to break even. A simple way to compare is return per unit of drawdown: Bot A earns about 2.2% APY for every 1% of drawdown, Bot B about 0.9%.

Neither bot is “better” in general. It depends on how large a fall you can accept. But you can only make that choice if you see all three numbers.

Questions to ask of any performance figure

1. What period does it cover? Exact start and end dates. Longer periods that include both rising and falling markets tell you more.
2. Is it net of fees? Performance fees, trading fees and funding can take a large share of gross returns.
3. Live money or backtest? A backtest shows how rules would have performed on past data. It has never faced real fills or slippage.
4. Realised only, or including open positions? Unrealised gains can disappear.
5. What was the maximum drawdown, and when? If it is not shown, ask.
6. Can you verify it? On Hyperliquid, a real track record has an address you can look up. See how to verify trading bot fills and PnL.

Reading CryptoBoost's numbers

The CryptoBoost homepage shows the strategy at about +22% APY, net of the 10% performance fee, with an account value chart from January 2024 to August 2026. Apply the same questions: look at the dips on the chart as well as the headline, and remember that yield is variable and not guaranteed. Past performance is not a guarantee of future results.

For your own account, Hyperliquid's Portfolio page shows your PnL and maximum drawdown for the period you select.

Common questions