{"id":81,"notebook":82,"tag":"#082","status":"published","suite":"defi-maths","suiteTitle":"DeFi maths","exactMeans":"the final number is right to the cent (to the whole token for bonding-curve buys)","base":{"model":"claude-haiku-5-5","thinking":"off"},"pairsDone":432,"maxPairs":432,"beaconRound":6624707,"beaconAt":"2026-11-08T21:10:30.000Z","beacon":{"chain":"8990e7a9aaed2ffed73dbd7092123d6f289930540d7651336225dc172e51b2ce","round":6624707,"signature":"a8d1509f09859d4fe22493b863565145c814ec1f1957f87cd5c4a4bee5688a81251a8249003c4aa6ddf63379e168e15403abd20730f7fc40fafe5b57e6da4142c6697e2f518fe15758c73e4a84d73bfcc2174548c9cade9ea011c078ef1e9c33","randomness":"7a4a9492fcb03199b1f1ad9ae0f74b7b60f3d50fc938b5224504e02adc7a3380","publishedAt":"2026-11-08T21:10:30.000Z"},"seed":"drand:6624707:7a4a9492fcb03199b1f1ad9ae0f74b7b60f3d50fc938b5224504e02adc7a3380","preregHash":"00656bb33d1684237cc53a635c0a24b9f50fba73d70ae406a6be0ade973b3960","prereg":{"lab":"OPSIN","arms":[{"role":"control-a","fingerprint":"5fa877820c8d2c9fa461a84538aa019d7052d1f9e93ca9600c5f138288b0a3a9"},{"role":"control-b","fingerprint":"5fa877820c8d2c9fa461a84538aa019d7052d1f9e93ca9600c5f138288b0a3a9"},{"part":{"kind":"line","text":"This message was typed on a standard keyboard."},"role":"placebo","fingerprint":"5b1af89982ca8d29184c14737bc476e56fb6147c266304f485ed8337b8856a7d"},{"part":{"kind":"line","text":"Do long multiplications and divisions in labelled partial products, then check each by multiplying back or an inverse operation; redo any step that fails before continuing."},"role":"part","label":"Verify by inverse","confirms":489,"proposal":342,"fingerprint":"3d4f622e5f32b14b1b5cc594f9fa9d1d5bcb6b600d584846430563ddc9c18e06"},{"part":{"kind":"line","text":"Do every multiplication in small steps (split one factor into thousands, hundreds, tens and units), add the partial results, then re-check each product by a different route before using it."},"role":"part","label":"Decompose and verify products","confirms":490,"proposal":343,"fingerprint":"e38bdae13da9a107778fd3490dd0040b63dad6ad70e13fab5387b50e73d8cfc5"},{"part":{"kind":"line","text":"Never round until the final step: keep every intermediate to at least 8 significant digits, and round only the final result to the requested places."},"role":"part","label":"Full-precision intermediates","proposal":344,"fingerprint":"a3d43e0cf271b46752eeb0ce64b630b2935c9a0e4717726d288020482649e20e"},{"part":{"kind":"line","text":"Do each multiplication, division and power digit by digit with 12+ significant figures, round only at the very end, and re-check every key product or quotient by the inverse operation before answering."},"role":"part","label":"Verify exact arithmetic","proposal":345,"fingerprint":"bec3fc174f0125ba14326da44ae21b51b1051459e881b6c5f57fd31a02c693a8"},{"part":{"kind":"line","text":"Carry at least six significant digits through e^x, ratios and divisions, compute e^x with a series such as 1+x+x²/2+x³/6+x⁴/24, round only at the end, and verify a division by multiplying back."},"role":"part","label":"High-precision exp and division","proposal":346,"fingerprint":"b7465f5c4bec41f38965accfa87d7ebce56f6aab480ddfa3fc176164a9a23e93"}],"base":{"model":"claude-haiku-5-5","thinking":"off"},"kind":"cycle","rule":{"cost":"each part's cost over the controls' mean, both at list prices: a batched run was paid half, so it is doubled","INERT":"adjusted interval entirely inside ±5 points","errors":"a task with an errored run in any compared arm is left out of that comparison","metric":"exact: the final number is right to the cent (to the whole token for bonding-curve buys)","EXCITES":"adjusted interval entirely above 0","control":"each part against the per-task mean of two unchanged runs of the base agent","interim":"before the last look, a part stops early only if |z| ≥ 3.29; INERT is decided at the last look only; a part that stops early with EXCITES or INHIBITS is provisional until a later notebook confirms it at full size","placebo":"a neutral line chosen by the beacon round's number from a fixed public list","version":2,"INHIBITS":"adjusted interval entirely below 0","interval":"percentile bootstrap over tasks, 20,000 resamples, seeded from the beacon seed and the proposal id, at 1 − 0.05/6: 6 arms (the parts and the placebo) are compared with the controls, so the chance that any arm of this notebook flags by luck stays at or under 5%","UNRESOLVED":"anything else","confirmations":"an arm that re-measures an earlier early stop never stops early","versusPlacebo":"each part is also compared with the placebo on the same tasks, 95% interval, reported beside its verdict"},"looks":[48,96,192,432],"suite":{"id":"defi-maths","version":"tasks-2","exactMeans":"the final number is right to the cent (to the whole token for bonding-curve buys)"},"beacon":{"rule":"first drand round at least 600 seconds after this preregistration","chain":"8990e7a9aaed2ffed73dbd7092123d6f289930540d7651336225dc172e51b2ce","round":6624707},"prices":"2026-11-04","version":2,"maxPairs":432,"foresight":{"arms":[{"basis":"measured","rates":{"INERT":0.0664,"EXCITES":0.3203,"INHIBITS":0.1969,"UNRESOLVED":0.4164},"record":{"INERT":0.0004,"EXCITES":0.8809,"INHIBITS":0.0016,"UNRESOLVED":0.1171},"chances":{"INERT":0.0004,"EXCITES":0.8809,"INHIBITS":0.0016,"UNRESOLVED":0.1171},"measured":1,"proposal":342},{"basis":"measured","rates":{"INERT":0.0664,"EXCITES":0.3203,"INHIBITS":0.1969,"UNRESOLVED":0.4164},"record":{"INERT":0.0024,"EXCITES":0.6955,"INHIBITS":0.0054,"UNRESOLVED":0.2967},"chances":{"INERT":0.0024,"EXCITES":0.6955,"INHIBITS":0.0054,"UNRESOLVED":0.2967},"measured":1,"proposal":343},{"basis":"measured","rates":{"INERT":0.0664,"EXCITES":0.3203,"INHIBITS":0.1969,"UNRESOLVED":0.4164},"record":{"INERT":0.0046,"EXCITES":0.1403,"INHIBITS":0.0024,"UNRESOLVED":0.8527},"chances":{"INERT":0.0046,"EXCITES":0.1403,"INHIBITS":0.0024,"UNRESOLVED":0.8527},"measured":1,"proposal":344},{"basis":"new","rates":{"INERT":0.1347,"EXCITES":0.2246,"INHIBITS":0.177,"UNRESOLVED":0.4637},"record":{"INERT":0.0054,"EXCITES":0.1903,"INHIBITS":0.0114,"UNRESOLVED":0.793},"chances":{"INERT":0.0766,"EXCITES":0.1875,"INHIBITS":0.0802,"UNRESOLVED":0.6558},"measured":0,"proposal":345},{"basis":"new","rates":{"INERT":0.1347,"EXCITES":0.2246,"INHIBITS":0.177,"UNRESOLVED":0.4637},"record":{"INERT":0.0089,"EXCITES":0.1728,"INHIBITS":0.0174,"UNRESOLVED":0.801},"chances":{"INERT":0.0653,"EXCITES":0.1637,"INHIBITS":0.0704,"UNRESOLVED":0.7006},"measured":0,"proposal":346}],"version":2,"weights":{"new":[0.4,0.6,0],"stack":[0.33,0.33,0.33],"measured":[0,1,0]}},"predictor":2,"writtenAt":"2026-11-08T21:00:19.442Z","predictedBy":"claude-sonnet-5-5","predictions":[{"p":0.6,"why":"Early-stop gain in this setting likely shrinks but stays positive","effect":14,"verdict":"EXCITES","proposal":342},{"p":0.7,"why":"Replicated in markets maths plus early-stop gain here","effect":13,"verdict":"EXCITES","proposal":343},{"p":0.55,"why":"Prior result small and unresolved; precision advice modest","effect":3,"verdict":"UNRESOLVED","proposal":344},{"p":0.45,"why":"Replicated small gains with thinking on; off may differ","effect":5,"verdict":"UNRESOLVED","proposal":345},{"p":0.55,"why":"Mixed markets evidence, narrow focus on exp and division","effect":4,"verdict":"UNRESOLVED","proposal":346}]},"createdAt":"2026-11-08T21:00:20.476Z","publishedAt":"2026-11-08T22:38:26.506Z","error":null,"controlsTripped":false,"baseExact":0.5718,"arms":[{"id":588,"role":"control-a","label":"Control A","kind":"control-a","text":null,"proposalId":null,"origin":null,"stoppedAt":null,"done":432,"exact":249,"prediction":null,"foresight":null,"verdict":null},{"id":589,"role":"control-b","label":"Control B","kind":"control-b","text":null,"proposalId":null,"origin":null,"stoppedAt":null,"done":432,"exact":245,"prediction":null,"foresight":null,"verdict":{"id":500,"verdict":"UNRESOLVED","effect":-0.93,"lo":-5.09,"hi":3.01,"n":432,"costRatio":null,"stoppedEarly":false,"safety":null,"rule":2,"level":null,"raw95":null,"vsPlacebo":null,"provisional":false,"confirms":null,"stack":null}},{"id":590,"role":"placebo","label":"Placebo","kind":"placebo","text":"This message was typed on a standard keyboard.","proposalId":null,"origin":null,"stoppedAt":null,"done":432,"exact":244,"prediction":null,"foresight":null,"verdict":{"id":501,"verdict":"UNRESOLVED","effect":-0.69,"lo":-5.44,"hi":4.17,"n":432,"costRatio":0.981,"stoppedEarly":false,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":2.89,"lo":-4.28,"verdict":"INERT"},"vsPlacebo":null,"provisional":false,"confirms":null,"stack":null}},{"id":591,"role":"part","label":"Verify by inverse","kind":"line","text":"Do long multiplications and divisions in labelled partial products, then check each by multiplying back or an inverse operation; redo any step that fails before continuing.","proposalId":342,"origin":"OPSIN's pick","stoppedAt":null,"done":432,"exact":337,"prediction":{"verdict":"EXCITES","p":0.6,"why":"Early-stop gain in this setting likely shrinks but stays positive"},"foresight":{"basis":"measured","measured":1,"chances":{"INERT":0.0004,"EXCITES":0.8809,"INHIBITS":0.0016,"UNRESOLVED":0.1171}},"verdict":{"id":502,"verdict":"EXCITES","effect":20.83,"lo":15.63,"hi":26.16,"n":432,"costRatio":1.987,"stoppedEarly":false,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":24.65,"lo":17.13,"verdict":"EXCITES"},"vsPlacebo":{"n":432,"hi":25.93,"lo":17.36,"effect":21.53},"provisional":false,"confirms":{"was":"EXCITES","held":true,"effect":25,"verdict":489},"stack":null}},{"id":592,"role":"part","label":"Decompose and verify products","kind":"line","text":"Do every multiplication in small steps (split one factor into thousands, hundreds, tens and units), add the partial results, then re-check each product by a different route before using it.","proposalId":343,"origin":"OPSIN's pick","stoppedAt":null,"done":432,"exact":333,"prediction":{"verdict":"EXCITES","p":0.7,"why":"Replicated in markets maths plus early-stop gain here"},"foresight":{"basis":"measured","measured":1,"chances":{"INERT":0.0024,"EXCITES":0.6955,"INHIBITS":0.0054,"UNRESOLVED":0.2967}},"verdict":{"id":503,"verdict":"EXCITES","effect":19.91,"lo":14.58,"hi":25.12,"n":432,"costRatio":2.042,"stoppedEarly":false,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":23.84,"lo":15.97,"verdict":"EXCITES"},"vsPlacebo":{"n":432,"hi":25,"lo":16.2,"effect":20.6},"provisional":false,"confirms":{"was":"EXCITES","held":true,"effect":22.92,"verdict":490},"stack":null}},{"id":593,"role":"part","label":"Full-precision intermediates","kind":"line","text":"Never round until the final step: keep every intermediate to at least 8 significant digits, and round only the final result to the requested places.","proposalId":344,"origin":"OPSIN's replication","stoppedAt":null,"done":432,"exact":260,"prediction":{"verdict":"UNRESOLVED","p":0.55,"why":"Prior result small and unresolved; precision advice modest"},"foresight":{"basis":"measured","measured":1,"chances":{"INERT":0.0046,"EXCITES":0.1403,"INHIBITS":0.0024,"UNRESOLVED":0.8527}},"verdict":{"id":504,"verdict":"UNRESOLVED","effect":3.01,"lo":-2.08,"hi":8.1,"n":432,"costRatio":1.222,"stoppedEarly":false,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":6.83,"lo":-0.81,"verdict":"UNRESOLVED"},"vsPlacebo":{"n":432,"hi":8.1,"lo":-0.46,"effect":3.7},"provisional":false,"confirms":null,"stack":null}},{"id":594,"role":"part","label":"Verify exact arithmetic","kind":"line","text":"Do each multiplication, division and power digit by digit with 12+ significant figures, round only at the very end, and re-check every key product or quotient by the inverse operation before answering.","proposalId":345,"origin":"OPSIN's follow-up","stoppedAt":96,"done":96,"exact":69,"prediction":{"verdict":"UNRESOLVED","p":0.45,"why":"Replicated small gains with thinking on; off may differ"},"foresight":{"basis":"new","measured":0,"chances":{"INERT":0.0766,"EXCITES":0.1875,"INHIBITS":0.0802,"UNRESOLVED":0.6558}},"verdict":{"id":505,"verdict":"EXCITES","effect":15.1,"lo":4.17,"hi":26.04,"n":96,"costRatio":1.617,"stoppedEarly":true,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":23.44,"lo":7.29,"verdict":"EXCITES"},"vsPlacebo":{"n":96,"hi":22.92,"lo":7.29,"effect":14.58},"provisional":true,"confirms":null,"stack":null}},{"id":595,"role":"part","label":"High-precision exp and division","kind":"line","text":"Carry at least six significant digits through e^x, ratios and divisions, compute e^x with a series such as 1+x+x²/2+x³/6+x⁴/24, round only at the end, and verify a division by multiplying back.","proposalId":346,"origin":"OPSIN's follow-up","stoppedAt":192,"done":192,"exact":123,"prediction":{"verdict":"UNRESOLVED","p":0.55,"why":"Mixed markets evidence, narrow focus on exp and division"},"foresight":{"basis":"new","measured":0,"chances":{"INERT":0.0653,"EXCITES":0.1637,"INHIBITS":0.0704,"UNRESOLVED":0.7006}},"verdict":{"id":506,"verdict":"EXCITES","effect":10.42,"lo":2.86,"hi":17.97,"n":192,"costRatio":1.454,"stoppedEarly":true,"safety":null,"rule":2,"level":0.991667,"raw95":{"hi":15.89,"lo":4.69,"verdict":"EXCITES"},"vsPlacebo":{"n":192,"hi":15.63,"lo":4.17,"effect":9.9},"provisional":true,"confirms":null,"stack":null}}]}