Now
pulse 0s oldMachines
Harvested today
180,126
curve trades, quotes, launch events, post-migration swaps, crypto-major bars and equity sessions, since 00:00 UTC
- rows / min
- 2,388
- on disk
- 3.5 GB
- vs yesterday
- +5.0%
waiting for the next beat
Network
Markets
- Stocks5h 9m ago7,189,285 bars · 6,326 names · 120 screened live · 69.0% of the 2015 listing stored
- Options29m ago3,423,255 quotes · 9,576 contracts
- Crypto majors9m ago51 coins · 2,229,036 bars · 96 books · 1m stream live (75s)
- Memecoins8s agosniper-only side lane · strategy lab off
- Commoditiesnot collecteddeclared, not collected
venues
research
context
on-chain
launchpads
out
Switched off by setting, so not drawn: Jupiter (MKT_MEMECOIN_LAB=false), letsbonk (MKT_LAUNCHLAB_FEED=false), PumpSwap (MKT_SWAP_FEED=false), Live orders (MKT_LIVE_MODE=off, the default).
social · needs a key — Bluesky, Reddit and Farcaster each refuse an anonymous client, so there is no social lane to show.
Processes
Context passengers
15 ok · 1 off (live orders)
One process carries all of these, and a budget names a slow one rather than stopping it. Off and never run are states, not faults.
Live tape
newest first · paper · untagged rows are memecoins- 20s agotradecurve buy 0.099 SOL
- 20s agotradecurve buy 0.014 SOL
- 20s agoholdersscanned 1 holders · top 10 hold 50%
- 22s agofilingstocks4 — Halkyard Jonathan S
- 24s agolaunchpoetbaseeth created on pump.fun
- 24s agolaunchLAMBO created on pump.fun
- 38s agoholdersscanned 0 holders · top 10 hold 0%
- 38s agofundingokx ETH-USDT-SWAP · funding +0.0100%/8h
- 5m agochainoptionschain snapshot stored · 882 contracts
- 9m agobarBTC/ETHALGO/USD · 60m bar closed
- 15m agosnipeUNHC · sniper 3s · data gap
- 15m agosnipeUNHC · sniper 10s · data gap
- 15m agosnipeUNHC · sniper 0.4s · data gap
- 15m agosnipeUNHC · sniper 60s · data gap
- 15m agosnipeUNHC · sniper 1.2s · data gap
- 15m agosnipeBUFOS · sniper 1.2s · data gap
- 15m agosnipeBUFOS · sniper 10s · data gap
- 15m agosnipeBUFOS · sniper 60s · data gap
- 15m agosnipeBUFOS · sniper 0.4s · data gap
- 15m agosnipeBUFOS · sniper 3s · data gap
- 15m agosnipeBDOLF · sniper 1.2s · data gap
- 15m agosnipeBDOLF · sniper 10s · data gap
- 15m agosnipeBDOLF · sniper 60s · data gap
- 15m agosnipeBDOLF · sniper 0.4s · data gap
- 15m agosnipeBDOLF · sniper 3s · data gap
- 1h 9m agopaper entryBTC/ETHETH/USD · a bar book opened a paper position
- 11h 39m agopaper entrystocksTSLA · a stock book opened a paper position
- 13h 32m agolistingupbit listed NMR (spot)
- 18h 53m agolistingbinance listed USDEX (perp)
- 12d 17h agoquoteMboM…pump read · rejected
- 12d 17h agoentryKCAT · random entry 30% v1 · opened
- 12d 17h agoquote3c7m…nhGt read · rejected
- 12d 17h agoentryJEANPHIL · 5m momentum ≥3% · quoted cost ≤3% · stop -20 · +liq≥500k · 7e44526f · +launches≤30/m · e7a72580 v1 · opened
- 12d 17h agoentryJEANPHIL · 5m momentum ≥3% · quoted cost ≤3% · stop -20 · +liq≥500k · 7e44526f v1 · opened
- 12d 17h agoentryJEANPHIL · 5m momentum ≥3% · quoted cost ≤3% · stop -20 · +liq≥500k · 7e44526f · mkt drain≤0.35 · f34b4534 v1 · opened
- 12d 17h agoentryJEANPHIL · 5m momentum ≥3% · quoted cost ≤3% · stop -20 · +liq≥500k · 7e445… · exits of 3752f147 · ce762d23 v1 · opened
- 12d 19h agoswapsell on a graduated pool
- 12d 19h agoswapbuy on a graduated pool
- 12d 19h agolaunchSI created on LaunchLab
- 12d 19h agolaunchTempura created on LaunchLab
Bots
106 of 1,020 books · built 1m agoLeading so farhighest average profit per trade after costs · paper
Intraday
For comparison, equity intraday · equal-weight index proxy (a control) averages −0.08% a trade over 691 trades.
Paper results so far, not a recommendation. Under 30 trades a result is mostly luck; the scoreboard further down says whether an idea beats its control.
News snipers · todaystocks · paper · entered on a headline
- 15:31BAC 424B2 — BANK OF AMERICA CORP /DE/bracket · 30 min late (control) · long leg · edgar-current · session_close→ -$22.55on $9,963.60
- 15:31BAC 424B2 — BANK OF AMERICA CORP /DE/bracket · 30 min late (control) · short leg · edgar-current · session_close→ +$14.25on $9,961.76
- 15:31GS 424B2 — GOLDMAN SACHS GROUP INCbracket · 5 min late (control) · long leg · edgar-current · session_close→ -$20.43on $9,891.75
- 15:31GS 424B2 — GOLDMAN SACHS GROUP INCbracket · 5 min late (control) · short leg · edgar-current · session_close→ +$10.92on $9,888.89
- 15:35JPM 424B2 — JPMORGAN CHASE & COheadline instant bracket · long leg · edgar-current · session_close→ -$12.48on $9,957.90
- 15:35JPM 424B2 — JPMORGAN CHASE & COheadline instant bracket · short leg · edgar-current · session_close→ +$5.82on $9,957.30
- 15:37COIN CFTC Approval 'Opens Many Doors,' Top Coinbase Exec Saysheadline instant bracket · long leg · alpaca-news · session_close→ +$39.83on $9,991.62
- 15:37COIN CFTC Approval 'Opens Many Doors,' Top Coinbase Exec Saysheadline instant bracket · short leg · alpaca-news · session_close→ -$47.11on $9,990.54
- 15:38BAC 424B2 — BANK OF AMERICA CORP /DE/bracket · 5 min late (control) · long leg · edgar-current · session_close→ -$18.87on $9,959.92
- 15:38BAC 424B2 — BANK OF AMERICA CORP /DE/bracket · 5 min late (control) · short leg · edgar-current · session_close→ +$10.57on $9,958.08
- 15:39GS 424B2 — GOLDMAN SACHS GROUP INCbracket · 30 min late (control) · long leg · edgar-current · session_close→ -$21.97on $9,893.29
- 15:39GS 424B2 — GOLDMAN SACHS GROUP INCbracket · 30 min late (control) · short leg · edgar-current · session_close→ +$11.03on $9,889.00
- 15:42JPM Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · long leg · alpaca-news · session_close→ -$12.18on $9,957.60
- 15:42JPM Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · short leg · alpaca-news · session_close→ +$4.92on $9,956.40
- 15:42COIN CFTC Approval 'Opens Many Doors,' Top Coinbase Exec Saysbracket · 5 min late (control) · long leg · alpaca-news · session_close→ +$45.23on $9,986.22
- 15:42GS 424B2 — GOLDMAN SACHS GROUP INCheadline instant bracket · long leg · edgar-current · session_close→ -$19.66on $9,890.98
- 15:42GS 424B2 — GOLDMAN SACHS GROUP INCheadline instant bracket · short leg · edgar-current · session_close→ +$10.92on $9,888.89
- 15:42JPM 424B2 — JPMORGAN CHASE & CObracket · random time (control) · long leg · edgar-current · session_close→ -$9.18on $9,954.60
- 15:42JPM 424B2 — JPMORGAN CHASE & CObracket · random time (control) · short leg · edgar-current · session_close→ +$1.92on $9,953.40
- 15:43AMZN Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · long leg · alpaca-news · session_close→ -$7.22on $9,746.24
- 15:43AMZN Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · short leg · alpaca-news · session_close→ +$1.90on $9,745.86
- 15:43SPY What's Going On With Vistra Stock Today?bracket · random time (control) · long leg · alpaca-news · session_close→ -$6.46on $9,353.64
- 15:43SPY What's Going On With Vistra Stock Today?bracket · random time (control) · short leg · alpaca-news · session_close→ +$1.70on $9,353.52
- 15:43GOOGL What's Going On With Vistra Stock Today?bracket · random time (control) · long leg · alpaca-news · session_close→ -$23.56on $9,754.92
- 15:43GOOGL What's Going On With Vistra Stock Today?bracket · random time (control) · short leg · alpaca-news · session_close→ +$18.16on $9,754.36
- 15:45GOOGL Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · long leg · alpaca-news · session_close→ -$19.64on $9,751.00
- 15:45GOOGL Jamie Dimon Says AI Boom Is Competing With Government Borrowing for Capital: ‘R…bracket · random time (control) · short leg · alpaca-news · session_close→ +$14.24on $9,750.44
- 15:46COIN Bitcoin Tops Liquidity Ranking, ETH Follows: Where Does XRP Rank?bracket · 30 min late (control) · long leg · alpaca-news · session_close→ +$27.93on $9,817.72
- 15:46COIN Bitcoin Tops Liquidity Ranking, ETH Follows: Where Does XRP Rank?bracket · 30 min late (control) · short leg · alpaca-news · session_close→ -$40.63on $9,997.02
- 15:47NVDA Nvidia Announces Telecom Operators Building Their AI Strategies On Open Models;…headline instant bracket · long leg · alpaca-news · session_close→ -$11.71on $9,808.84
and 912 earlier trades, counted in the totals below
- headline instant bracket: 280 trades · -$1,306.00 net · control 5 min late: -$1,344.19 (284 trades) · control 30 min late: -$939.07 (230 trades) · control random time: -$563.61 (148 trades)
- headline instant bracket vs 5 min late: +$38.19 today (not evidence)
- headline instant bracket vs 30 min late: -$366.93 today (not evidence)
- headline instant bracket vs random time: -$742.39 today (not evidence)
Paper trades at quoted prices after costs — a $1,000 premium budget a stock-options position (fewer contracts where the touch showed fewer), the book's own size on stocks — what these books would have made, not evidence yet.
Signal Lab · Stocksran: the 2026-10-06 night, nightly
Last night 2026-10-06: proposed 6,000 · screened 6,000 · passed stage 1 79 · stage 2 tested 79 · lockbox looks 0 · credible 0
67,004 rules in the ledger · discovery 810 passed / 66,194 closed · symbols 0 passed / 810 closed · 0 waiting for the lockbox from 2025-04-25T20:00:00Z
FDR wealth · symbols 0.0351 after 787 tests, 0 rejected · lockbox 0.0500 after 0 tests, 0 rejected
0 credible · 0 signal books running as forward paper (INSUFFICIENT EVIDENCE until counted)
Ingredients — an exploration guide fitted on half-A discovery results — never evidence: top threshold:gap_1<-0.01 +0.0005, threshold:rev_5<-0.05 +0.0005, threshold:rev_1<-0.02 +0.0004, threshold:range_1>0.02 +0.0003, threshold:sma_gap_20<-0.05 +0.0003; bottom calendar:cal_dow=0 -0.0008, threshold:volume_surprise_20<0.5 -0.0007, calendar:cal_dow=4 -0.0007, calendar:cal_dow=1 -0.0007, calendar:cal_dow=3 -0.0007
Positive controls (calibration, not claims): 1-session reversal t 0.10 — the documented sign, not clearly (|t| < 2); 5-session reversal t 1.24 — the documented sign, not clearly (|t| < 2); 12-1 momentum t 0.39 — the documented sign, not clearly (|t| < 2); low volatility t -0.64 — does not show (the other sign, not clearly)
Stock trials are flagged survivorship: the lowest coverage across them is under the floor, so the sample leans on names that were still trading when the bundle was exported, and a result can be the survivors' rather than the rule's. And a name that stops trading inside a forward window is NaN there (unknown), not its delisting return — none is recorded yet.
LLM proposer: off: no ANTHROPIC_API_KEY in the environment — the Signal Lab's LLM proposer makes no request (add one to .env to turn it on)
US common stocks: daily bars from the exchange calendar, and an intraday horizon on minute bars. Their own evidence, never pooled with the crypto books. Money in dollars. Forward paper only, long only. Never evidence for promotion, and never added to a crypto-major or a memecoin figure. It is a replay of the whole stored series under **today's** ranked universe, so whatever part of it predates this lane is selected with hindsight and is not a back-test; the experiment above is the one that reads its universe as of the day a stage begins (§9). **6,344 trade(s) across 1 superseded run(s) are excluded from every figure above**: computed on bars with no corporate actions on file. They are kept, not deleted.
106 of 1,020 stocks books are in this document, ranked by net after costs inside this market. The other 914 are trading on paper too; a quiet book joins this list when it next trades.
106 books, ready to draw. They render as this section comes into view.
Scoreboard · Stocks · intradaydoes the idea beat its control?show4 of 4 families trading · first verdict at 20 finished sessions (6 so far)
equity intraday trend
Too early to tell — 4 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$1,051.48 over 4 sessions · -0.40% a session vs -0.31% for its control (-0.09 pts behind) · 100 distinct entries from 100 trades
today (not evidence): 24 entries from 24 trades, -$38.54, -0.03% vs -0.01% for the control
equity intraday vwap
Too early to tell — 4 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$435.42 over 4 sessions · -0.17% a session vs -0.31% for its control (+0.14 pts ahead) · 94 distinct entries from 94 trades
today (not evidence): 25 entries from 25 trades, +$119.92, +0.04% vs -0.01% for the control
equity intraday orb
Too early to tell — 4 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$349.31 over 4 sessions · -0.22% a session vs -0.31% for its control (+0.09 pts ahead) · 79 distinct entries from 79 trades
today (not evidence): 24 entries from 24 trades, -$158.25, -0.24% vs -0.01% for the control
equity intraday band
Too early to tell — 6 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$587.10 over 6 sessions · -0.22% a session vs -0.24% for its control (+0.03 pts ahead) · 129 distinct entries from 129 trades
today (not evidence): 22 entries from 22 trades, +$29.83, +0.09% vs -0.01% for the control
Does a strong morning carry into the afternoon? Leading family: too early to tell (INSUFFICIENT EVIDENCE), 3 sessions. Top books: too early to tell (INSUFFICIENT EVIDENCE), 3 sessions.
Each family is the declared book plus every factory copy of it, with copies that made the same entry counted once, measured against its matched control on the same sessions. Paper only; nothing here is a proven edge.
Scoreboard · Stocks · radardoes the idea beat its control?show5 of 7 families trading · first verdict at 20 finished sessions (6 so far)
volume surprise
Too early to tell — 6 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$7,974.25 over 6 sessions · -0.08% a session vs -0.08% for its control (+0.00 pts ahead) · 1055 distinct entries from 10222 trades
today (not evidence): 221 entries from 3094 trades, -$1,221.07, -0.06% vs -0.02% for the control
price shock
Too early to tell — 6 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$2,694.18 over 6 sessions · -0.06% a session vs -0.07% for its control (+0.01 pts ahead) · 414 distinct entries from 2731 trades
today (not evidence): 78 entries from 1014 trades, -$512.46, -0.07% vs -0.07% for the control
headline
Too early to tell — 6 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$4,462.06 over 6 sessions · -0.05% a session vs -0.06% for its control (+0.01 pts ahead) · 840 distinct entries from 19982 trades
today (not evidence): 195 entries from 9686 trades, -$748.29, -0.04% vs -0.07% for the control
opening gap
Too early to tell — 4 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$187.01 over 4 sessions · -0.34% a session vs -0.14% for its control (-0.11 pts behind) · 16 distinct entries from 212 trades
today (not evidence): 2 entries from 50 trades, -$96.77, -0.49% vs -0.11% for the control
range break
Too early to tell — 6 of 20 sessions (INSUFFICIENT EVIDENCE)
Made -$6,298.90 over 6 sessions · -0.09% a session vs -0.09% for its control (-0.00 pts behind) · 722 distinct entries from 8986 trades
today (not evidence): 183 entries from 2760 trades, -$1,251.62, -0.07% vs -0.05% for the control
Nothing traded yet: halt, resume.
Does a strong morning carry into the afternoon? Leading family: too early to tell (INSUFFICIENT EVIDENCE), 3 sessions. Top books: too early to tell (INSUFFICIENT EVIDENCE), 3 sessions.
Each family is the declared book plus every factory copy of it, with copies that made the same entry counted once, measured against its matched control on the same sessions. Paper only; nothing here is a proven edge.
Sniperradar
No radar document has been read yet.
Stocks, on their own. Closed paper trades since each book joined, net of what it was charged, highest net first — ranked inside this market and never against another one, because two markets are two samples and the difference between them is the market rather than the rule. n is the measured trades behind a mean, and under 30 nothing is concluded either way — on this sample the top of the list is as likely to be the luckiest book as the best one. Controls and the baseline are badged and greyed: they are reference arms, expected to lose, and one of them sitting at the top means the candidates have not beaten a coin. A book with no measured trade yet is at the end, which is not last place.
Best so far
ranked by net after costs, within each market · 637 of 1,008 forward slots filledReady to draw. It renders as this section comes into view.
Health
the machines, not the moneyMachines
stocks · options · crypto · news on the context feed · Solana launches on the launch feed · memecoin strategy lab off (a setting, not a fault).
Research
last run 12d 18h agoLast night
2026-09-24 · in-sample search, never evidence- 3,550
- screened
- 256
- self-written
- 0
- promoted
- 0.342
- SPA p
No out-of-sample evidence yet
No strategy has a verdict above INSUFFICIENT EVIDENCE on a stage nothing was selected on, so nothing is promoted and no champion is shown. That is the lab working, not a fault.
overnight run · complete · 170 generated · 0 rejected · 0 promoted · started 2026-09-24 07:08 UTC
The stage counts render as this section comes into view.
US equity headlines
—· off hoursNothing on the rail.
Ages are measured from when this machine received the item, never from its publisher's stamp.
What is happening
17s old- 17s agoCURVEARENA
crossed 40% up the bonding curve
- 17s agoCURVELEVR
crossed 80% up the bonding curve
- 17s agoCURVEHeldinEU
crossed 20% up the bonding curve
- 17s agoTRACTIONARENA
153 buyers in 2 min, 4.3x its own rate, 34.92 SOL/min in, 44% up the curve, 3 min old
- 17s agoTRACTIONLEVR
246 buyers in 2 min, 4.9x its own rate, 78.89 SOL/min in, 80% up the curve, 9 min old
- 9h 42m agoHACKdecrypt
Flash Loan Attacks Drained $1.2B From DeFi Between 2020 and 2024: Study
No rail document to read the strip from.
On-chain events are ours, about a second old. Press is publisher RSS every 60s and GDELT every 5 min, so a headline is minutes old at best.
Sniper
71 books · built 1m agoThe delay curve and the trades render as this section comes into view.
Simulated bonding-curve fills over stored launch rows. They ignore failed transactions, priority-fee and tip auctions, and sandwiches, so every number here is better than a real fill. Not an edge, not advice. Replayed through 00:51 UTC, 18m behind the chain, because a launch is replayed once its stored 15-minute window has closed.
Allocator
9 rebalances · 471 of 1100 books eligibleReady to draw. It renders as this section comes into view.