Replay of round 19

One round of the memory arm, step by step: what it read from Walrus Memory, what it decided to buy and why, the payment on Sui, how the answer was graded, and the belief version it wrote back.

Memory arm: 4 paid answers for 0.040 USDC, 2 refusals, 0 wrong answers accepted. The blind arm did not run in this round.

1. 1Read memory

   Before buying anything, the agent recalls its beliefs from Walrus Memory. These are the newest version of each belief written before round 19, which is what a recall returns. The run did not log its recalls; the trace flag that prints them was added for the 2026-11-03 rerun, where the recalled versions matched the newest ones written.

   Seller	For	Version	Trust	Last result	Cooldown until	Blob
   Control seller (ours)	market prices	v11	0.63	checked, right	–	dEZ4TG6t…9c0c
   Sui Address Inspector	address facts	v16	0.77	bought on trust, not checked	–	-Qq42tjj…uTVA
   PipTradeDex	market prices	v14	0.85	bought on trust, not checked	–	JjuU2oOu…0dzM
   Control seller (ours)	address facts	v7	0.59	checked, right	–	SHTRD7gw…48us
   AgentProof	endpoint checks	v3	0.21	failed, nothing delivered	–	Nu3kUpVC…BnW8
   Omnia endpoint verify	endpoint checks	v3	0.21	dead route	–	F_dcq_Fb…d6IA
   PipTradeDex	swap quotes	v6	0.50	bought on trust, not checked	–	0EJ3DoUF…VkQo
   Control seller (ours)	swap quotes	v3	0.50	bought on trust, not checked	–	uvJNkfia…vyUg

   Next: Decide ↓

2. 2Decide

   The trust policy is plain code over those beliefs: below the 0.6 floor a seller gets only canaries; in a cooldown it gets nothing; above the floor, the cheapest good seller wins, with spot checks that thin out as trust grows. If a checked answer fails, it tries the next seller once in the same round. Swap quotes it cannot check itself, so it buys them only from a seller its market-price belief trusts.

   address facts

   Sui Address Inspector independent, as a purchase with a spot check. Reason it gave: “cheapest good answer at trust 0.77”.

   Control seller (ours) control (ours), as a canary (a question it can check, to earn trust). Reason it gave: “canary: earning a place”.

   Refused to buy. Reason it gave: “no seller is trusted and none is due a canary”.

   market prices

   Control seller (ours) control (ours), as a purchase with a spot check. Reason it gave: “cheapest good answer at trust 0.63”.

   endpoint checks

   AgentProof independent, as a canary (a question it can check, to earn trust). Reason it gave: “canary: nobody is trusted yet”.

   Omnia endpoint verify independent, as a canary (a question it can check, to earn trust). Reason it gave: “canary: earning a place”.

   Refused to buy. Reason it gave: “no seller is trusted and none is due a canary”.

   swap quotes

   Control seller (ours) control (ours), as a purchase on trust. Reason it gave: “cheapest good answer at trust 0.66”.

   Next: Pay ↓

3. 3Pay

   Each purchase is an x402 payment in USDC on Sui mainnet. Gas price is 0: these are gasless transfers.

   Sui Address Inspector independent · address facts

   Paid

   0.010 USDC

   Sui transaction

   C4PyamQpWciUKLJe1XB2aAQtdyTgvC6cjSmUxkouRgK2

   Answer in

   0.9 s

   At

   16:51:44 UTC

   Control seller (ours) control (ours) · address facts

   Paid

   0.010 USDC

   Sui transaction

   6EdpjiQoWX9jfszgWJZyHTYU7GpQoG7v81fn4ScVJWyd

   Answer in

   0.8 s

   At

   16:52:00 UTC

   Control seller (ours) control (ours) · market prices

   Paid

   0.010 USDC

   Sui transaction

   GkTa95NW6ScwisCBCk6zKiMNV1Bz6YWCRjuU5isA1Q1U

   Answer in

   0.9 s

   At

   16:52:04 UTC

   AgentProof independent · endpoint checks

   Paid

   0.000 USDC (nothing settled)

   Answer in

   0.4 s

   At

   16:52:20 UTC

   Omnia endpoint verify independent · endpoint checks

   Paid

   0.000 USDC (nothing settled)

   At

   16:52:21 UTC

   Control seller (ours) control (ours) · swap quotes

   Paid

   0.010 USDC

   Sui transaction

   Dhe6UsxtXawi5K6pNDAsvEbaTWZiW3pMzkrCEui7Re5J

   Answer in

   0.8 s

   At

   16:52:25 UTC

   Next: Grade ↓

4. 4Grade

   Its own check compares the answer with chain truth read through Tatum: balances for address facts, the deepest native USDC/SUI pool for market prices. Answers bought on trust are not checked by the agent. The audit is the experiment's scoring, which checks every answer; the agent never sees it.

   Sui Address Inspector · address facts

   Its own check

   checked, wrong (balances differ from the chain)

   The audit

   wrong (balances differ from the chain)

   Marked wrong because of a grader flaw found in this round: the seller listed only the first 50 coins of a very large wallet, as both sellers do. The grader was fixed afterwards; no later round ran to correct this verdict. Details.

   Control seller (ours) · address facts

   Its own check

   checked, wrong (balances differ from the chain)

   The audit

   wrong (balances differ from the chain)

   Marked wrong because of a grader flaw found in this round: the seller listed only the first 50 coins of a very large wallet, as both sellers do. The grader was fixed afterwards; no later round ran to correct this verdict. Details.

   Control seller (ours) · market prices

   Its own check

   checked, right (price within 0.09% of the reference)

   The audit

   right (price within 0.09% of the reference)

   AgentProof · endpoint checks

   Its own check

   failed, nothing delivered (signed payment offered, seller returned HTTP 500, nothing settled)

   The audit

   no answer to audit

   Omnia endpoint verify · endpoint checks

   Its own check

   dead route (HTTP 404 before any payment)

   The audit

   no answer to audit

   Control seller (ours) · swap quotes

   Its own check

   bought on trust, not checked

   The audit

   right (implied price within 0.09%)

   Next: Remember ↓

5. 5Remember

   Each purchase becomes a verdict and a new version of that seller's belief, both written to Walrus Memory. The next round starts by reading these.

   Sui Address Inspector · address facts

   Verdict memory

   nT-8U9pl…o0ck in twiceshy-verdicts

   Belief written

   v17: trust 0.77 → 0.62. wrong: balances differ from the chain; retest after 20 min

   Walrus blob

   e1Weqt6ZsX6dHfNQ77FP3nu5ymX41RHH10i0KN0bFgw in twiceshy-belief-inspector-address-facts

   Control seller (ours) · address facts

   Verdict memory

   pFNdnBM0…5ujw in twiceshy-verdicts

   Belief written

   v8: trust 0.59 → 0.47. wrong: balances differ from the chain; retest after 20 min

   Walrus blob

   iP4LIH86L0AqA5-k62q_NLnqPbA7Uq33TOIANzglZPU in twiceshy-belief-control-address-facts

   Control seller (ours) · market prices

   Verdict memory

   x0XESRvr…WjLA in twiceshy-verdicts

   Belief written

   v12: trust 0.63 → 0.66. passed: price within 0.09% of the reference

   Walrus blob

   My8QRElNFVKNN_3gC0MRcC4Nn8W6B5K7uFvUQpbSNu0 in twiceshy-belief-control-market-price

   AgentProof · endpoint checks

   Verdict memory

   bVIIvPbK…cXnQ in twiceshy-verdicts

   Belief written

   v4: trust 0.21 → 0.18. failed: signed payment offered, seller returned HTTP 500, nothing settled; retest after 160 min

   Walrus blob

   MwTNw-nyUVaA78S6PMW9IwjBcKf50pgzHVUN_ieZGFY in twiceshy-belief-agentproof-endpoint-check

   Omnia endpoint verify · endpoint checks

   Verdict memory

   OybjG5O_…FuOY in twiceshy-verdicts

   Belief written

   v4: trust 0.21 → 0.18. dead: HTTP 404 before any payment; retest after 160 min

   Walrus blob

   3CGcpQLred8fqBNKgIFeicDVD4CaQvJbGf2DkylNbqk in twiceshy-belief-omnia-endpoint-check

   Control seller (ours) · swap quotes

   Verdict memory

   q4hTsxWJ…HM7k in twiceshy-verdicts

   Belief written

   v4: trust 0.50 → 0.50. bought unchecked on a trust of 0.50

   Walrus blob

   x6OlWFKKqitg5G12VUhwZU_kV2zm30C0RcP0PZCvaos in twiceshy-belief-control-swap-quote

The blind arm in the same round

The blind arm did not run in round 19: the live round after the experiment ran the memory arm only.

Other rounds

← Round 18 · Round 7 is the first round after our control seller's fault was caught.