Skip to content

FAQ

Questions people ask before signing up and in the first few weeks after. If something is going wrong right now, start with Troubleshooting; if a term is unfamiliar, see the Glossary.

AlphaHouse runs a quantitative, market-neutral trading system inside your own exchange account. You keep your money where it is: we connect over an API key, compute a target book each day, and place the orders that move your account toward it.

The strategy trades crypto perpetual futures and holds roughly twenty positions at a time, split between longs and shorts. It aims to earn from how those assets perform relative to each other rather than from the market going up. See What AlphaHouse is, and Target exposure for how the book is sized.

How is this different from other trading bots?

Section titled “How is this different from other trading bots?”

Most retail bots are single-instrument rules: a grid, a moving-average cross, a DCA ladder. One decision at a time, one asset.

AlphaHouse works at the portfolio level. Every day it ranks a large universe of perpetual futures, builds a target weight for each name, nets that against what you already own, and sends only the difference. Positions are sized against your equity and leverage setting, capped per symbol so no single name can dominate, and rebalanced on a schedule rather than on a trigger.

The book holds longs and shorts at roughly equal size, so the two sides move together when the whole market moves. What is left is the difference between them. On Portfolio that shows up as a net exposure figure near zero alongside a much larger gross exposure figure, which is the healthy state rather than a bug.

Market-neutral is not the same as risk-free. See “What if I lose money?” below.

No. Your funds never leave your exchange account, and no code path anywhere in the platform initiates a withdrawal or transfer.

The part you control is the permission scope on the API key: grant trade permissions and leave withdrawal permissions off. The exchange enforces that, which is why it is the layer worth getting right. Setup has the exact toggles per venue.

On Hyperliquid the guarantee is stronger still: an API wallet cannot sign a withdrawal at the protocol level, whatever permissions you think you granted. See Hyperliquid.

Bitget, Gate, and Hyperliquid. Each has its own credential shape and minimum:

Exchange Credentials needed Minimum account
Bitget API key, API secret, passphrase $2,500
Gate API key, API secret $2,500
Hyperliquid Wallet address, API wallet private key $5,000

The minimums come from each venue’s minimum order size, not from policy.

We add venues deliberately, selecting for breadth of listed markets and depth of liquidity, because a long-tail strategy needs both. If the venue you want is not here yet, sign up and name it at the connect-exchange step: those requests set the order we work in. See Choosing a venue.

$2,500 on Bitget or Gate, about $5,000 on Hyperliquid.

Arithmetic rather than a rule we chose. The strategy holds around twenty positions, so each leg is roughly a twentieth of your gross book, and every exchange refuses orders below a minimum notional. Below these figures a real share of your legs fall under that minimum and are never placed.

Those legs appear as held orders, so the consequence is visible rather than hidden. A small account does not break, it just runs a coarser approximation of the book.

No. There is no performance fee and no profit share. You pay a flat subscription based on the size of book you run, and nothing else.

That is deliberately unlike exchange copy-trading and on-chain vaults, which commonly take 10 to 20 percent of profits. The fee is capped and predictable: it does not grow with returns. Honest in the other direction too, since in a flat or losing year you pay what you would pay in a good one.

How much does the subscription cost as a percentage of my account?

Section titled “How much does the subscription cost as a percentage of my account?”

At the top of each tier’s recommended account range it works out to roughly 5 to 6 percent of account equity per year on annual billing.

We would rather state that plainly than have you work it out afterwards. There is no performance fee on top, and the ranges are set so the percentage stays roughly level as you move up. At the bottom of a range it is higher, which is the honest argument for picking the tier that matches your account.

Yes, with no long-term commitment. Two things happen on different clocks:

  • Revoking access is immediate. Delete the API key at your exchange, or the connection under Accounts, and no further orders can be placed.
  • Billing stops at the end of the current period. Cancelling in Settings schedules it. You keep access until the period you have paid for ends, and the app shows the date. You can reactivate before then.

Neither closes your positions. If you want to be flat, close them yourself at the exchange.

Yes, with a nuance we insist on stating.

This is not a curve-fitted backtest. The signals have been generated and recorded as they happened for over two years, with no lookahead: each day’s ranking was written down before the market traded it. Think of it as live paper trading that never stopped, with real capital in the market for the recent stretch. Published results are net of modelled fees, slippage, and funding, and every drawdown is on the performance page.

The caveats: the risk overlays sitting on top of the core ranking were developed later against that recorded history, so they are the fitted part and we say so on the performance page. And past performance does not guarantee future results.

Losing stretches are normal and should be expected. Market-neutral positioning reduces dependence on market direction, it does not remove risk: both books can move against you, funding can run against the book, and leverage magnifies each of those. We publish every drawdown and make no claims about not losing. Only trade with capital you can afford to lose.

If you are in a losing stretch right now and deciding what to do about it, read The mental game before you touch anything. The two reactions that feel most natural are the two that reliably cost money.

What if the strategy loses money in my first few months?

Section titled “What if the strategy loses money in my first few months?”

If the strategy’s net return in your account is negative after your first 90 days, contact us and we will refund the subscription fees you paid for that period.

To be explicit about the boundary: we refund what you paid us. Not trading losses, exchange fees, funding payments or slippage, because that money went to the market and to your exchange rather than to us.

Four conditions apply, worth knowing at the start rather than at the end:

  • Your account must have stayed connected and enabled for substantially all of the 90 days. Disconnecting, pausing, trading manually in the same account, or closing positions by hand all void it, because the recorded return no longer reflects the strategy.
  • Net return is measured excluding your own deposits and withdrawals, from the trading activity recorded in your account.
  • Request it in writing within 30 days of the 90-day period ending.
  • Once per customer.

The authoritative wording is in the Terms. If you intend to rely on this, read that section before you pause anything.

Four steps, covered in full in the setup walkthrough:

  1. Create an API key at your exchange with trade but no withdrawal permission.
  2. Add the connection under Accounts and run Test Connection.
  3. Choose a system and a plan.
  4. Set your leverage and start the system.

After that it runs on a schedule. Your first week covers what to expect, including the parts that look wrong but are not.

Usually because it was too small for the exchange to accept, and the system declined to send it rather than have it rejected. Every exchange enforces a minimum order size and a quantity step, and a rebalance worth a few dollars on one name can fall under it. Rounding up would mean trading a size the model did not ask for.

Orders that were held covers each reason, including the ones that do need your attention.

Most likely because our risk rules classified the market as unfavourable and the system reduced gross exposure. It is a deliberate risk control, applied to the whole book at once, and it reverses when conditions do. See Regime scaling, and Troubleshooting to confirm that is what happened rather than something else.

Why do the numbers on Dashboard and Portfolio disagree?

Section titled “Why do the numbers on Dashboard and Portfolio disagree?”

They usually measure different things: one aggregates across every connected exchange, the other shows a single connection, and the two pages use different time windows. See the entry in Troubleshooting.

You can, but the system will trade against you. It computes the difference between your actual positions and its target book, so a manual position in a name it holds looks like drift and gets rebalanced away, and one in a name it does not hold may be closed as an unwanted holding. Use a separate exchange sub-account instead.