trader-agent v85

Evaluation

v85paper window 2026-09-02 10:16Z → now, 0.1 d resolved 2 / 40 retired: superseded-same-day

sim

+2803.2 /day prod, same window

+5256.2 /day

HOLD noise floor ±306.4/day (5 pts of prod stake) · markets bet by both 9

window 0.1d < 7d
resolved 2 < 40
sim earns +2803.2 mana/day (de-scaled /1) vs prod +5256.2, beyond the 5.0pt-of-stake noise floor (306.4/day)

Modules of v85

Own trades and own declines only; each module's page has the reasoning.

codex-openlive

+646 / +648

A price-blind language-model estimate on each new market; bets only when it disagrees with the price by more than the configured edge.

0bets by entry price1

bets: 8
staked: 456
settled: 5/6 won
realised: +617
open (2) sold: +29
open (2) held: +32
return sold: +142%
return held: +142%

down-fadelive

+171 / +171

0bets by entry price1

bets: 2
staked: 45
settled: 2/2 won
realised: +171
open (0) sold: +0
open (0) held: +0
return sold: +379%
return held: +379%

close-longlive

+2 / +2

Longshots at 0.03-0.15 inside 24h of close resolve NO about 95% of the time.

0bets by entry price1

bets: 2
staked: 50
settled: 2/2 won
realised: +2
open (0) sold: +0
open (0) held: +0
return sold: +4%
return held: +4%

close-favlive

+2 / +2

Favourites at 0.90-0.97 inside 24h of close resolve YES about 99% of the time.

0bets by entry price1

bets: 1
staked: 25
settled: 1/1 won
realised: +2
open (0) sold: +0
open (0) held: +0
return sold: +7%
return held: +7%

whale-fadelive

+0 / +0

A single 100+ mana YES bet that pushes a mid-life market up 5+ points is overconfident on average; fades it.

0bets by entry price1

bets: 3
staked: 0
settled: none settled yet
realised: +0
open (0) sold: +0
open (0) held: +0
return sold: +0%
return held: +0%

class-fadelive

-10 / -7

0bets by entry price1

bets: 2
staked: 50
settled: none settled yet
realised: +0
open (2) sold: -10
open (2) held: -7
return sold: -20%
return held: -14%
booked off probe: 1

open-snipelive

-17 / -17

Bets a brand-new untraded market from its opening price toward the historical resolution rate of its cell (type x price x horizon x ante).

0bets by entry price1

bets: 1
staked: 17
settled: 0/1 won
realised: -17
open (0) sold: +0
open (0) held: +0
return sold: -100%
return held: -100%

market-keeperlive

-26 / +90

0bets by entry price1

bets: 2
staked: 206
settled: none settled yet
realised: +0
open (2) sold: -26
open (2) held: +90
return sold: -13%
return held: +44%

codex-swingoff

no fills

Exit policy for codex-open positions.

disabled in this version's config

event-makeroff

no fills

Quotes both sides of busy event markets and collects the spread.

disabled in this version's config

mc-indeplive

no fills

Independent multiple-choice answers listed at creation resolve YES far less often than their 0.50 opens imply.

judged 107 markets, placed nothing

mc-otherlive

no fills

The catch-all 'Other' answer in winner-take-all markets is overpriced above 0.20; buys NO toward 0.12.

judged 107 markets, placed nothing

numeric-tailshadow

no fills

Buys the lowest bucket of a new numeric or date market.

judged 107 markets, placed nothing

park-nolive

no fills

Parks idle mana as NO in stale mid-priced books instead of leaving it in cash.

judged 7 markets, placed nothing

Trades of v85

when	module	market	bet	result	why
09-02 13:08	whale-fade	Will the Republicans Gain Maine´s 2nd Congressional District in 2026	NO 0	unfilled	fade: single 64-mana YES bet pushed 0.59->0.77; big YES aggression is overconfident on average
09-02 13:02	codex-open	Important Bayesian updating market ×3 answers	YES 190	settled +46	luna-medium price-blind says 0.45 vs answer 0.33 (edge +0.12 >= 0.1): Base rate is 1/3 (equal type sampling; no historical series exists); the first measurement favors an isosceles triangle but ordering rules are ambiguous, so I pull toward 0.5, with four further daily measurements before the Sep 9 close expected from the user-provided stream as the resolver and no matching witness market identified.
09-02 12:47	down-fade	Will GPT-6 be called "Astra"?	YES 18	settled +17	down-fade: single 70-mana NO bet pushed 0.76->0.39; big NO aggression is overconfident on average
09-02 12:46	down-fade	Is OpenAI's upcoming "Astra" model going to be called GPT-6?	YES 28	settled +153	down-fade: single 1000-mana NO bet pushed 0.35->0.15; big NO aggression is overconfident on average
09-02 11:32	codex-open	Заскамит? ×2 answers	NO 200	settled +570	luna-medium price-blind says 0.50 vs answer 0.80 (edge -0.30 >= 0.1): Без описания неясно, кто должен стримить, что считается скамом и какой источник разрешает исход, поэтому базовую частоту оценить нельзя и вероятность смещена к 0.5.
09-02 11:28	whale-fade	Will September 2026 headline CPI-U rise at least 0.3% month-over-month? ×2 answers	NO 0	unfilled	fade: single 100-mana YES bet pushed 0.63->0.78; big YES aggression is overconfident on average
09-02 11:21	codex-open	Will Grok 4.7 be released on 9/11?	NO 1	settled +1	luna-medium price-blind says 0.20 vs market 0.43 (edge -0.23 >= 0.2): Base rate for release on any one day in a ten-day window is about 10%; Musk’s September 2 announcement points to September 12, not 11, and absent resolution rules create date, timezone, and release-surface ambiguity, so I estimate 20%, with the official xAI News/Models pages as resolver and no matching witness market.
09-02 11:18	market-keeper	Will the top Artificial Analysis Intelligence Index score reach 65 or high ×2 answers	NO 206	open -26	keep: price 0.98, fv 0.91 (gap -0.072), target 103, held 0; sources calibrated 0.91
09-02 11:17	class-fade	Will Keir Starmer lose his seat at the next general election?	NO 25	open -21	class-fade: unseen account (no record) pushed YES 0.81->0.89 with 75 mana; such pushes fade
09-02 11:16	close-fav	Will Alcaraz win the next round at the US open? C.Alcaraz vs J.Faria	YES 25	settled +2	favorite band 0.90-0.97 within 24h of close resolves YES 99%
09-02 11:02	open-snipe	Will Alcaraz win the next round at the US open? C.Alcaraz vs J.Faria	NO 17	settled -17	virgin market: cell [sports\|hi\|<10d\|sm] historically resolves YES 0.42 vs creator open 0.89
09-02 10:54	class-fade	Will OpenAI officially confirm that Astra uses a recurrent-depth (looped)	NO 25	open +11	class-fade: unseen account (no record) pushed YES 0.28->0.56 with 50 mana; such pushes fade
09-02 10:17	codex-open	Who will win the Massachusetts governor’s race? ×2 answers	NO 65	open +29	luna-medium price-blind says 0.03 vs answer 0.50 (edge -0.47 >= 0.1): Base rate: Republicans won 3 of Massachusetts’ last 6 gubernatorial elections (50%), but Healey leads Minogue 54.3%–31.3% in the RCP average; the Nov. 3 election is the key scheduled event, and official certification by the Massachusetts Secretary of the Commonwealth will resolve the market, with the fixed-answer N/A clause adding only a negligible tail risk.
09-02 10:16	close-long	Will Julian Alvarez be sold to Barcelona by the end of the summer transfer	NO 25	settled +1	longshot band 0.03-0.15 within 24h of close resolves NO 95%
09-02 10:16	close-long	Will Dylan Matthew beat Community Prediction (Metaculus Summer Cup)	NO 25	settled +1	longshot band 0.03-0.15 within 24h of close resolves NO 95%

The lineage

version	mana	status · window	bets	staked	realised	open held	open sold	per day
v89	real	live in prod 2026-09-19 16:42Z → running	221	5,709	+338	+386	-116	+242.3
v88	real	retired, book still settling 2026-09-03 13:36Z → 2026-09-19 16:42Z	1117	30,685	+928	+4016	+1235	+306.5
v81	real	retired, book still settling 2026-08-30 09:55Z → 2026-09-03 13:36Z	393	13,334	+1646	+760	+5	+579.2
v80	real	retired, book still settling 2026-08-29 19:42Z → 2026-08-30 09:55Z	26	595	+7	-1	-6	+11.3
v78	real	retired, book still settling 2026-08-29 17:15Z → 2026-08-29 19:28Z	11	411	+25	-47	-71	-242.6
v77	real	retired, book still settling 2026-08-29 16:03Z → 2026-08-29 17:15Z	5	247	-18	-32	-52	-985.7
v77	real	retired, book still settling 2026-08-29 16:03Z → 2026-08-29 17:15Z	5	247	-18	-32	-52	-985.7
v76	real	retired, book still settling 2026-08-29 15:27Z → 2026-08-29 16:02Z	10	298	+320	+0	+0	+13115.0
v75	real	retired, book still settling 2026-08-29 13:01Z → 2026-08-29 15:27Z	13	218	+149	+0	+0	+1479.0
v74	real	retired, book still settling 2026-08-29 10:35Z → 2026-08-29 13:01Z	12	128	+23	-14	-22	+89.3
v73	real	retired, book still settling 2026-08-29 10:24Z → 2026-08-29 10:35Z	10	250	+3	+2	+1	+616.5
v72	real	retired, book still settling 2026-08-29 10:12Z → 2026-08-29 10:24Z	5	125	+2	+1	+1	+343.1
v71	real	retired, book still settling 2026-08-29 10:03Z → 2026-08-29 10:12Z	5	125	+2	+2	+2	+553.2
v70	real	retired, book still settling 2026-08-29 06:12Z → 2026-08-29 10:03Z	9	150	-9	+8	+8	-3.0
v67	real	retired, book still settling 2026-08-28 07:38Z → 2026-08-29 06:12Z	27	475	-108	+40	+21	-72.3
v66	real	retired, book still settling 2026-08-28 06:43Z → 2026-08-28 07:38Z	2	50	+1	+0	-0	+36.5
v65	real	retired, book still settling 2026-08-27 19:48Z → 2026-08-28 06:43Z	18	253	+340	+3	-3	+753.4
v64	real	retired, book still settling 2026-08-27 10:06Z → 2026-08-27 19:48Z	9	227	+36	-64	-72	-69.3
v63	real	retired, book still settling 2026-08-27 08:12Z → 2026-08-27 10:06Z	3	75	+5	-3	-5	+36.5
v62	real	retired, book still settling 2026-08-27 07:18Z → 2026-08-27 08:12Z	3	69	+5	+3	+2	+224.9
v89-twin	paper	evaluating 2026-09-19 16:43Z → running	196	3,632	+191	+1131	+520	+442.7
codex-paper	paper	retired: window-complete 2026-08-22 19:49Z → 2026-08-22 19:49Z	1	40	+0	+13	+4	-
v88-twin	paper	retired: prod-moved 2026-09-03 16:10Z → 2026-09-19 12:21Z	1412	26,687	+1027	+5778	+2442	+429.6
v87	paper	retired: superseded-same-day 2026-09-03 00:02Z → 2026-09-03 15:21Z	54	952	-375	+95	-2	-438.5
v86	paper	retired: window-complete 2026-09-02 13:27Z → 2026-09-09 15:32Z	539	13,928	+1183	+3446	+1059	+653.2
v85	paper	retired: superseded-same-day 2026-09-02 10:16Z → 2026-09-02 13:08Z	21	849	+775	+114	-7	+7461.3
v84	paper	retired: superseded-same-day 2026-09-02 02:00Z → 2026-09-02 08:22Z	46	350	-40	+42	+18	+5.7
v83	paper	retired: window-complete 2026-09-01 10:09Z → 2026-09-08 03:32Z	646	11,605	+727	+2860	+835	+533.5
v82	paper	retired: window-complete 2026-08-30 22:07Z → 2026-09-01 09:44Z	177	3,963	+247	+167	-2	+279.2
v80-twin	paper	retired: prod-moved 2026-08-29 22:05Z → 2026-08-30 07:42Z	28	650	+61	+43	+20	+259.9
v77-twin	paper	retired: prod-moved 2026-08-29 16:04Z → 2026-08-29 17:52Z	15	400	-116	+170	+54	+723.9
v75-twin	paper	retired: prod-moved 2026-08-29 13:31Z → 2026-08-29 15:38Z	16	474	-200	+35	-4	-1880.5
v72-twin	paper	retired: prod-moved 2026-08-29 10:14Z → 2026-08-29 12:19Z	11	250	+26	+20	+5	+532.6
v70-twin	paper	retired: prod-moved 2026-08-29 08:17Z → 2026-08-29 08:29Z	6	140	+2	+8	+8	+1241.1
v68	paper	retired: window-complete 2026-08-28 13:52Z → 2026-08-30 18:48Z	73	2,211	+410	+330	+249	+335.1
v67-twin	paper	retired: prod-moved 2026-08-28 08:16Z → 2026-08-29 07:19Z	32	792	+239	+112	+80	+365.9
v65-twin	paper	retired: prod-moved 2026-08-27 20:15Z → 2026-08-28 02:27Z	17	238	-57	+16	+10	-159.6
v64-twin	paper	retired: prod-moved 2026-08-27 10:07Z → 2026-08-27 20:07Z	13	334	+34	-51	-58	-41.4
v63-twin	paper	retired: prod-moved 2026-08-27 08:14Z → 2026-08-27 08:18Z	5	347	+3	+22	+5	+11001.8
v62-twin	paper	retired: prod-moved 2026-08-27 07:20Z → 2026-08-27 07:48Z	6	310	+3	+24	+13	+1336.9
v60-twin	paper	retired: prod-moved 2026-08-27 10:06Z → 2026-08-27 10:06Z	2	50	+3	+0	+0	+109728.0
v59-twin	paper	retired: prod-moved 2026-08-26 20:46Z → 2026-08-26 20:46Z	3	84	+0	+11	+1	+460637.2
v58-twin	paper	retired: prod-moved 2026-08-26 12:59Z → 2026-08-26 19:49Z	13	391	-67	+13	-11	-190.2
v57-twin	paper	retired: prod-moved 2026-08-26 12:24Z → 2026-08-26 12:51Z	2	8	+0	+12	+4	+626.0
v56-twin	paper	retired: prod-moved 2026-08-26 12:01Z → 2026-08-26 12:10Z	3	110	+0	+75	+49	+11809.4
v55-twin	paper	retired: prod-moved 2026-08-26 11:41Z → 2026-08-26 11:59Z	5	101	+1	+78	+55	+6374.5
v53-twin	paper	retired: prod-moved 2026-08-26 10:58Z → 2026-08-26 11:04Z	4	70	+22	+5	-2	+6307.7
v50-twin	paper	retired: prod-moved 2026-08-26 10:03Z → 2026-08-26 10:33Z	5	121	+34	+28	+21	+2885.8
v45	paper	retired: superseded-same-day 2026-08-25 20:32Z → 2026-08-25 21:04Z	8	605	-86	+8	-41	-3429.4
v44	paper	retired: superseded-same-day 2026-08-25 20:28Z → 2026-08-25 20:31Z	7	365	+15	-6	-32	+5636.0
v41	paper	retired: superseded-same-day 2026-08-25 14:46Z → 2026-08-25 17:58Z	9	656	-481	+3	-5	-3573.8
v39	paper	retired: superseded-same-day 2026-08-25 12:44Z → 2026-08-25 13:31Z	1	100	-100	+0	+0	-3092.3
v38	paper	retired: superseded-same-day 2026-08-25 12:18Z → 2026-08-25 12:41Z	3	212	-12	+27	+15	+973.2
v37	paper	retired: superseded-same-day 2026-08-25 09:45Z → 2026-08-25 09:45Z	1	28	+0	+14	+12	+37215.2
v35	paper	retired: superseded-same-day 2026-08-25 08:50Z → 2026-08-25 08:59Z	6	184	+57	+29	+23	+14065.5
v34	paper	retired: superseded-same-day 2026-08-25 08:46Z → 2026-08-25 08:47Z	1	131	+57	+0	+0	+258245.1
v20	paper	retired: superseded-same-day 2026-08-24 13:39Z → 2026-08-24 14:34Z	3	203	-10	+94	+90	+2187.4
v16	paper	retired: superseded-same-day 2026-08-24 12:42Z → 2026-08-24 13:15Z	15	375	-96	+4	+0	-4042.7
v15	paper	retired: superseded-same-day 2026-08-24 12:38Z → 2026-08-24 12:39Z	9	250	-77	-26	-26	-111952.5
v14	paper	retired: superseded-same-day 2026-08-24 12:35Z → 2026-08-24 12:36Z	7	222	-74	-4	-4	-86088.4
v13	paper	retired: superseded-same-day 2026-08-24 12:31Z → 2026-08-24 12:32Z	8	227	-71	-4	-4	-87939.7
v12	paper	retired: superseded-same-day 2026-08-24 12:16Z → 2026-08-24 12:17Z	8	222	-67	-4	-4	-92033.7
v11	paper	retired: window-complete 2026-08-23 20:07Z → 2026-08-30 04:03Z	359	10,830	+1707	+1482	+487	+503.8
v10	paper	retired: superseded-same-day 2026-08-23 08:06Z → 2026-08-23 20:03Z	30	907	+216	+257	+189	+951.2
v9	paper	retired: window-complete 2026-08-22 20:03Z → 2026-08-29 00:28Z	327	7,435	+2028	+972	+410	+485.1
v8	paper	retired: superseded-same-day 2026-08-22 17:52Z → 2026-08-22 19:52Z	16	400	+76	+8	+3	+1010.6
v7	paper	retired: superseded-same-day 2026-08-22 08:16Z → 2026-08-22 17:06Z	6	120	+96	+95	+33	+518.9