Hakaru (測る) — Japanese: “to measure”

Measured. Not predicted.

Hakaru is one trend-following strategy — KAI 潮流 — run on real capital across 36 markets and 5 asset classes. It doesn’t forecast; it measures which markets are trending, sizes each by its own volatility, and publishes everything: the live result, the drawdowns, the losing windows. A track record you can’t audit is marketing, not evidence.

Factsheet — KAI 潮流 · liveLive from the exchange, updating as the book trades.
KAI — realised P&L
Syncing
KAI — open P&L
Open positions
Days live

01 / Performance

The whole record. Not the good quarter.

10-year backtest · 36 markets

Annual return
+6.3%
Volatility
15.1%
Sharpe
0.41
Max drawdown
−25.4%
Last 180 days (paper)
−4.46%

Known limit, stated here and not in a footnote: perpetual funding is not yet simulated. On a 3× gross book held 7–14 days, funding is the same order of magnitude as the modelled return — until it lands in the backtest, +6.3% is an upper bound.

How to read this against alternatives

  • Managed-futures ETFs offer a similar mechanic in a brokerage account, with a decade of audited returns. If that fits you better, buy the ETF — KAI exists for people who want the mechanic inside their own exchange account, with every number auditable.
  • Exchange copy-trading shows self-reported leaderboards with unknown risk. KAI publishes its methodology, its limits and its losing windows instead.
  • BTC buy & hold is a different bet entirely — higher historical return, ~60% volatility, −80% drawdowns. KAI targets 15% vol and trades both directions.

02 / Methodology

The mechanic behind managed futures, on one account.

36 curated instruments — crypto, equity indices, commodities, single stocks, FX — traded long and short on perpetual futures. No indicators invented in a forum; no model that claims to know the future.

01

Measure the trend — never predict it

Momentum sign over 20/40/60/100-day lookbacks, averaged to [−1, 1]. When horizons disagree, the position shrinks. A weak trend earns a small position, not a full one.

02

Size by each market’s own volatility

Inverse-volatility weights: quiet markets get more notional, wild ones less. Every position risks a comparable amount.

03

Scale the whole book to one risk budget

The portfolio is scaled by its realised covariance to a 15% annualised volatility target. Caps bind after scaling — the book may fall short of target rather than pile into the calmest cluster.

04

Rebalance on a measured clock

Per-asset-class cadence, because the classes measured differently: crypto fast, TradFi slow. Delta rebalancing only — never flatten-and-reopen.

Caps: ±10% per instrument · 3.0× grossProtection: 12% catastrophe stop per position, resting at the exchangeExecution: delta rebalancing only — never flatten-and-reopen

03 / What we don’t promise

The section every trading product leaves out.

No win rate

KAI reports P&L in aggregate from the exchange’s own income history. A per-trade “win rate” would have to be invented — so it isn’t.

No prediction

In sideways markets the strategy loses money. That is the price of the trending ones, not a malfunction. It is stated here, not in a footnote.

No averaging into losses

Position size follows signal and volatility — never the wish to be right.

No hidden bad months

Live results come from the exchange, update as the book trades, and stay red when they are red.

Leveraged trading in perpetual futures can lose money fast, and past results — ours included — guarantee nothing about the future. The live track record is short. That is the most important limitation of all, which is why it is on this page.

04 / Security & custody

Your money never touches us.

Trade-only API keys

Withdrawal permission is never requested and never needed. Your funds cannot leave your exchange account through Hakaru.

Your account, your custody

Positions are mirrored onto your own BingX account. Hakaru never pools or holds customer funds.

Exchange-side catastrophe stop

A 12% protective stop per position rests at the exchange. It fires even if every Hakaru server is offline.

A kill switch you control

Pause new entries from the dashboard or the mobile view — or delete the API key and the connection is gone.

05 / Access

Watch first. Connect later.

Waitlist — now

We are opening access in a limited first batch. Join the waitlist and be first through the door when the next cohort opens.

Join the waitlist
Members — how it works
  • Trade-only BingX API keys — withdrawal permission never needed
  • KAI’s book is mirrored onto your account; per-member loss limits apply
  • Catastrophe stops rest at the exchange and fire even if we are offline
  • Pause everything with one switch; delete the key and we are gone

No guaranteed returns. You remain in control of your account and your risk.

06 / FAQ

Asked by sceptics. Answered the same way.

Why should I trust a strategy whose last 180 days are negative?+

You shouldn’t — not on that window alone. Trend following loses in sideways markets; six months is too short to conclude anything in either direction. That is why the live figures on this page come from the exchange and update as the book trades: the evidence is the page, not a promise.

How is this different from exchange copy-trading?+

Leaderboards show self-reported, often cherry-picked track records with unknown risk. KAI is one documented strategy: methodology, backtest limits, and losing windows are published. You can verify before you connect a key.

What happens to my money if Hakaru disappears tomorrow?+

Nothing. Funds never leave your exchange account; API keys are trade-only. Open positions keep their exchange-side catastrophe stops. Delete the key and the relationship is over.

Can I run my own trades alongside KAI?+

Yes. An ownership registry means KAI only ever closes what KAI opened. Your manual positions are listed separately and are untouchable by the engine.