Papertrade just told PAPER holders they won’t be able to sell their tokens anytime soon, and it did that on purpose.

On Monday, the team behind Hyperliquid’s new 1,000x perps exchange said it wasn’t planning on letting PAPER leave its staking contract anytime soon: “At this juncture we do not have plans to allow off-contract token transferability.”

— (@)

That’s a big turn from the original plan. When we previewed Papertrade’s launch earlier this month, PAPER was set to eventually become transferable, and the team’s Oct. 1st announcement post said transfers would come “after a sufficient period of time has passed.”

Now, that activation is off the table for the foreseeable future, and multiple venues like Aster and Paragon have delisted their pre-market PAPER perps. Plenty of the traders who’d been losing money on purpose to farm PAPER are, understandably, furious.

Whatever you think about this move, though, the reasoning behind the call has revealed something fundamental about this new style of “house pool” exchange. Namely, when a token is a claim on the house, shorting that token is a bet against the house, and that bet can pay for an attack on it.

Papertrade’s tough decision

First, a quick refresher for anyone who missed our recent article. Papertrade itself acts as the counterparty to every trade via one shared USDC pool, which fills up from traders’ losses.

When you lose, you’re minted PAPER, and staking PAPER earns you USDC from the protocol’s take, including whatever spills over once the pool passes its cap. As we explained back in May, losers slowly end up owning a piece of the house.

Monday’s update changed a few things:

1. PAPER can still be staked and unstaked, but it can’t be sent to another wallet or sold, and the team has no plans to change that right now.

2. Going forward, every $1M in trader losses will mint ~0.1% of PAPER’s supply at that moment, tilting new issuance toward newer users and slowly diluting older holders. Papertrade also cut rewards: At launch, every $1 you lost earned up to 100 PAPER. Now it’s about 10. The exception is if the pool gets close to running out of money: then a bonus kicks in to pull in more losing trades. The team calls this “the house running a discounted auction on its own future cash flows.”

These changes will be queued in a timelock and go live 24 hours later, pending final security checks. So why now?

The aforementioned Paragon listed its PAPER perp on Oct. 18th, just eight days after launch, and Aster followed suit on Monday morning. As co-founder Jez admitted afterward, “[I] specifically underestimated the demand for pre-market perps.”

Tradable PAPER as a weapon

PAPER’s yield comes from the house’s earnings, so its value rises and falls with the health of the pool. Drain the pool, or leave winners stuck in its payout queue (i.e. the line where unpaid winnings wait on fresh losses), and PAPER will crater.

That’s where the trouble would start. Papertrade marks its trades to Hyperliquid’s midpoint (i.e. halfway between the top bid and ask), so briefly shoving Hyperliquid’s price shoves Papertrade’s too.

For example, on launch night one wallet nudged ETH ~0.12% on Hyperliquid and came away about $1.28M ahead, per The Block, though Papertrade says the trader later gave most of it back through new accounts.

The team has been candid on this vector. In an Oct. 11th note, it wrote that “a manipulation attack is always possible — whether economically advantageous to attempt is the question.”

The default answer to that has been to keep the math unfavorable, partly by capping how big the bankroll gets. A smaller pool means a smaller prize. Adding a PAPER short to the mix changes things, though.

Here’s the team’s own example. Say the pool holds $10M. Moving BTC or ETH prices enough to drain it would cost an attacker about $15M, so on its own the attack loses money. As the team put it, “You simply lose money.” Now say that attacker is also short $10M of PAPER on another exchange. Draining the pool would crash PAPER, the short would pay off, and that covers the cost. Suddenly the attack is “very viable indeed.”

In Papertrade’s words, a PAPER perp turns into “a leveraged bet on bankroll solvency.”

The problem is that those shorts would sit on other exchanges, which Papertrade doesn’t control. It can limit the size of its own pool, but it can’t stop Paragon, Aster or anyone else from listing PAPER markets. Its only defense would be to limit how much people can trade on Papertrade itself. So every new PAPER market elsewhere would put a cap on how big Papertrade can grow.

Forced to choose between a liquid token and a growing exchange, it picked its exchange.

Curiously, this blind spot shows up in the best prelaunch stress test I saw. In August, Sumcap replayed ~3.6M Hyperliquid trades through Papertrade’s design across 600+ scenarios, with the pool staying solvent in every one. The catch is that Sumcap’s model explicitly left PAPER’s market price out, which is where this attack would live.

What now?

So what does a PAPER holder own today? In short, a share of the protocol’s USDC payouts that you can collect, yet the shares themselves can’t be cashed out. Jez said it’s “better to think of [PAPER] as a revenue share in the exchange than a traditional crypto token.”

Those payouts aren’t trivial, to be sure. When I pulled Papertrade’s numbers from DefiLlama, it had sent roughly $58M to PAPER holders since launching on Oct. 10th. They’ve also cooled considerably, though, from a peak of ~$6.7M on Oct. 15th to an average of ~$1.5M a day over Oct. 17–19.

Meanwhile, the deal for new losers has also gotten worse.

At recent payouts of ~$0.00017 per token per day, the 10 PAPER from a $1 loss earn about $0.002 a day, so it’d take nearly 600 days just to earn back the $1. Even the biggest farmers aren’t ahead yet, either. Per Lookonchain, 21 of the 25 largest losers were still in the red after counting their staking rewards.

Moreover, the last price the market put on PAPER pointed the same way. Paragon settled its perp at $0.026, which values the ~8.85B PAPER supply at roughly $230M. If that recent ~$1.5M daily pace held for a full year, stakers would collect ~$560M, more than double that. Traders clearly didn’t expect it to hold.

This dynamic isn’t lost on the team. Monday’s post described the overhaul as turning PAPER back into “something much closer to the exchange rakeback it was intended to be,” i.e. a cut of the house’s take handed back to the players who fund it.

In other words, PAPER is becoming less of an asset and more of a loss rebate paid out in slow motion.

The other questions

Who’s going to keep losing?

Every dollar paid to stakers comes from trader losses. Plenty of farmers lost on purpose betting that PAPER would eventually be sellable.

Without an exit, farming becomes a ~1.6-year wait (nearly 600 days) just to break even, and that’s if payouts don’t keep shrinking. If those farmers drift away, what’s left is organic losing from people who’d gamble at 1,000x anyway. That’s Papertrade’s real revenue base, and we’re about to learn how big it is.

Can the shorts be stopped?

To an extent, sure. Paragon’s perp only made sense on the expectation that PAPER would trade someday, and with no future spot price to converge to, Paragon shut it down. Side doors remain, though.

PaperDAO’s PULP token, which trades on Uniswap, is a pro rata claim on a treasury that farms PAPER, yet its docs only promise redemption into PAPER “once Papertrade enables transfers.” Tradable wrappers like that are tiny today, but they’re the kind of exposure Papertrade can’t cap.

Can any house pool have a liquid token?

Jupiter’s YOLO, which launched Oct. 15th with the same payout math, skipped a token entirely. Its pool, which Jupiter calls the House, is funded by losses, walled off from Jupiter’s main JLP pool, and there’s simply nothing to short.

Cardboard, a newer Solana venue offering up to 1,000x against a shared house balance, took Papertrade’s route instead. It mints its CARDBOARD reward token at 100 per $1 of eligible losses, and its docs say tokens can’t move “until the protocol enables transfers.”

Per Pine Analytics, it drew ~$4.48B in opening notional from Oct. 15–19, a sizable share of it from matched long/short pairs that looked a lot like farming. Cardboard hasn’t publicly addressed Papertrade’s transferability call, but its farmers now have a fresh precedent to price in.

Is this permanent?

The “At this juncture” wording leaves the door open, clearly. Papertrade concedes its design is only as good as the price data Hyperliquid feeds it, and it’s betting that Hyperliquid’s order books will keep getting deeper.

That matters because deeper books make manipulation costlier, which could someday make a liquid PAPER safe. A price feed that’s harder to push around than a live midpoint would help, too, with YOLO’s delayed oracle fills as one example. Until then, it’s safe to assume PAPER will stay put.

Zooming out

Papertrade’s call will cost it lots of goodwill, of course. People who farmed PAPER expecting to sell it now hold a yield stream they can’t exit, plus the new minting curve will thin that stream over time. For many users, that’s not the deal they thought they signed up for.

Still, I think the team made a defensible choice. The pool exists to pay winners, so a token that makes attacking the pool profitable puts every trader at risk, not just stakers.

The bigger lesson as I see it is for the whole genre. In a house pool exchange, the token is the house, and letting the house’s equity trade freely hands attackers a hedge. Expect every house pool launch from here on, Cardboard included, to face this same obstacle.

The house still wins, sure, but Papertrade also just made it a lot harder to bet against. Keep this one on your radar, as there’s surely more intrigue to come here.