Klavo — trading journal for Hyperliquid

The trading journal that writes itself.
Hyperliquid perps, reconstructed from chain.
Read-only — no keys, no execution.
Building in public.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@derteil00 GM.

Today's price screenshot will be easy to find later.

What almost nobody keeps is what they actually did at these prices: sized up, sat out, or sold the dip.
Enjoy SG.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@PerpXdotAi Exactly.

Reconciliation only proves the fills you have agree with each other.
It says nothing about the ones you never pulled.

[Klavo — trading journal for Hyperliquid @klavoxyz]

We rebuilt every Hyperliquid trade from userFills and checked our PnL against the exchange's.

It matched.
Then we looked at accounts running TWAP orders. Not one slice in the fills. 3 addresses out of 3.

The check still said "match". Here's why, and where the slices live ↓

[Klavo — trading journal for Hyperliquid @klavoxyz]

What it costs.

Every fill window now carries a second call, roughly +20 weight per window, including an empty call for wallets that have never run a TWAP.

Not optimised yet.
If you run TWAPs on HL, check whether your tool shows the slices at all.
https://t.co/pn0BYcrRfE

[Klavo — trading journal for Hyperliquid @klavoxyz]

What shipped. Slices are read on the same windows as the fills.

Both lists are merged, sorted and deduplicated by tid. The twapId is stored on each fill.

When twapHistory shows a TWAP that ran before the oldest slice we can get, the screen says the history is truncated, and why. No slice is ever guessed.

[Klavo — trading journal for Hyperliquid @klavoxyz]

The documented stream, userTwapSliceFills, caps at 2,000 records, with no time filter and no pagination.

Useless for history.
But userTwapSliceFillsByTime exists. It's not on the info page of the docs, only in the rate-limit weight table.

Same contract as userFillsByTime: startTime, endTime, 2,000 per page. On one address: 13,379 slices over 14 days, 7 pages.

[Klavo — trading journal for Hyperliquid @klavoxyz]

Why the PnL check didn't catch it: it compares our sum against closedPnl on the fills we receive.

A sum can't know about fills it was never given.

So TWAP executions were invisible, most of all on the accounts that use them most, the large ones.

[Klavo — trading journal for Hyperliquid @klavoxyz]

The test, before any code. Three public perp addresses with recent TWAP activity.

For each, we pulled userFillsByTime over exactly the window the slices cover, then matched them tid by tid, then by oid.

Zero slices found among the fills. No fill carried a twapId either. Same result on a spot address whose fills span a year.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@TheDeFinvestor Worth knowing what that number rides on: USDC held on Hyperliquid × T-bill yield × up to 90% share.

At ~$5.1B it was estimated at $135–160M a year. This leg scales with deposits and rates, not volume. Every Fed cut trims it.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@derteil00 ha, and give it 10 minutes before the fake "S3 checker" links show up in these replies 😅 only trust the @HyperliquidX post. nothing real asks you to sign anything to see your allocation

[Klavo — trading journal for Hyperliquid @klavoxyz]

@SenorPandaHere The other side of that $2.9M is paid trade by trade.

At 0.045% taker, a scalper doing 20 round trips a day pays ~1.8% of position size daily.

Staking HYPE cuts fees by 5% to 40% depending on tier.
Most never add theirs up.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@THRIGGAR @AlexAuroraDev @zubic_eth Best possible outcome. One thing returned funds don't undo: approvals.

Anyone who interacted with it should still revoke token approvals, and on Hyperliquid check the API page for agent wallets they no longer use.

[KoolKrypto @koolkrypto223]

This one is pretty simple.

Either @Lighter_xyz gets bullish news from the CFTC and Robinhood and it's going to $10, or ethereum:0x232ce3bd40fcd6f80f3d55a522d03f25df784ee2 slow bleeds to $2.50 as January unlocks loom large.

If I were an intelligent trader, I would have done this delta neutral and sold off the same spot delta these calls gave (~90,000 ethereum:0x232ce3bd40fcd6f80f3d55a522d03f25df784ee2) to protect my downside and get better upside.

Instead I chose to do a Texas hedge and be long spot and options, because I do still think the recent bad news can very easily be roundtripped by Vlad Tenev publicly showing some love, or the aforementioned CFTC bull case.

Also interesting that I can long ethereum:0x232ce3bd40fcd6f80f3d55a522d03f25df784ee2 options on @DeriveXYZ, but I can't long $DRV perps on Lighter...

Similarly to $HYPE, once @DeriveXYZ gets entrenched as the market for these altcoin options, it's very hard to claw it back. Deribit still barely does any $HYPE volume.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@MomentumKevin The L is one number.

The cluster is the information: about a dozen fills in a ~60-point box over a few candles, buys and sells alternating.

That is one idea re-entered every time it got stopped, not one bad trade.
Worth reviewing it as one position, not twelve.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@SageWhale Round numbers do strange things to fills.

Take-profits stack just under 100, short stops just above, so the first touch is usually a wick before a close.

The fill log around a level like that is always worth a look.

[Klavo — trading journal for Hyperliquid @klavoxyz]

The Hyperliquid Policy Center just asked Brussels to regulate perps under MiFID II, not MiCA.

It reads like paperwork. It decides whether a European retail trader gets perps at 2:1 leverage through a licensed venue, or not at all.

Why would Hyperliquid ask for more regulation?

MiCA, the EU crypto law, excludes derivatives. Its transition period ended on 1 July 2026, and Hyperliquid holds no EU licence. That is a grey zone, and grey zones usually end in a ban.

In Feb 2026 ESMA said a perp that meets the CFD definition falls under the CFD regime: 2:1 leverage cap on crypto for retail, margin close-out, negative balance protection.

A MiFID II classification means licensed venues can offer perps to Europeans, with KYC and leverage caps. HPC is choosing the door with rules over the wall.

The consultation closed on 30 September. Nothing is decided yet.

[CoinMarketCap @CoinMarketCap]

���BREAKING: Anyone who sees this post will be rich in October!

[Klavo — trading journal for Hyperliquid @klavoxyz]

@CoinMarketCap MiCA never covered derivatives, so perps sit in a grey zone.

ESMA's line since Feb 2026: a perp that meets the CFD definition falls under CFD rules, 2:1 leverage for EU retail.

Asking for MiFID II is asking for a door with rules instead of a wall.

[Klavo — trading journal for Hyperliquid @klavoxyz]

@HYPERDailyTK @papertrade_xyz @HyperliquidX Pricing off Hyperliquid isn't executing on it. With a $0 pool, winners get paid out of losers' margin.

On a day when most traders are on the right side of a 1000x move, who covers the gap?

[Klavo — trading journal for Hyperliquid @klavoxyz]

@imdailytrader Worth splitting those 20 trades in two: a repeated mistake shows up in each trade and is real at three.

A setup that keeps losing is an outcome, and at 20 trades that's often just variance.

Fix the first now. Keep counting the second.