GitHub: https://github.com/abrownfox0/abrownfox001-twap60-prediction-trigger-system
YouTube walkthrough: https://www.youtube.com/watch?v=XzhugRL6BV4
Part 3 inferred the signal constraints: thresholded P(up), official TWAP first, permission to decay.
Part 4 reconstructs the slot machine from timestamps. You do not need engine.ts to see the loop. The public fills already sketch it.
Live profile: @abrownfox001
1. The Lifecycle You Can Infer
SlotOpen
→ wait
→ BUY near 50¢
→ re-score
├─ decayed → SELL ~flat
└─ intact → hold → redeem $1
That is not branding. It is what a mid-band account with ~27% sells and later exit timing has to be doing.
Public code that implements the same shape: src/engine.ts + src/clob.ts.
2. Entry Timing Is the First Tell
Observed intra-slot activity:
| Seconds after open | Share |
|---|---|
| 0–5s | 6.3% |
| 5–30s | 16.6% |
| 30–150s | 35.7% |
| 150–270s | 26.1% |
| 270–300s | 15.3% |
Median buy ~47s.
Most work is not a t=0 sniper print.
Outside inference:
- the bot waits for a path, not the opening quote
- it sources 45–55¢ instead of lifting the first thin offer
- clips stay small (~$50–175) and capped per slot
If an account dumps size in the first 3 seconds every market, it is a different machine.
3. Exit Timing Is the Second Tell
Median sell ~226s.
Sells sit later than buys.
That split is the scratch signature:
- early: accumulate when the mid-band is still tradeable
- later: flatten tickets whose
P(up)no longer clears conviction
~930 sampled markets bought and sold at almost the same price (0.504 → 0.503).
Those are not take-profit scalps. They are flat exits.
A hold-all bot would not need that many same-market sells.
4. Re-Score Has to Exist
You cannot observe the 10–30s loop directly from one fill.
You can infer it:
- same market, buy then sell
- sell much later than buy
- spread ≈ 0
- 5-minute clock still open
That only happens if something inside the slot changed its mind. In this architecture that something is an updated TWAP60 P(up).
After the 30s → 60s cutover, the official path is smoother. The loop should still exist. The decay constants should not be copied from the snapshot era.
5. What Redeem Means in Public Data
The resolved-book sample is the hold path:
- 150 markets near 50¢
- held-side 60.7%
- that book is what survived scratching
So the outsider should split every account into two ledgers:
- Gross activity — all buys and sells
- Settlement book — tickets still open at resolution
If you judge the bot only on ledger 1, scratches look like noise.
If you judge it only on ledger 2, you miss the filter that created the 60.7%.
6. Execution Constraints Visible From Outside
At 50¢ the binding problems are not fee theater:
- fill quality
- latency
- depth for a true flat scratch
- no double-sends
- per-slot caps
A “breakeven scratch” on a one-cent-wide book is a process.
The same phrase on a four-cent hole is a loss.
The Windows binary and npm start dry-run exist so someone can inspect the shell. They do not prove live fill quality.
7. Falsifiers for the Lifecycle Story
The design is wrong if public timing shifts to:
- almost all fills in 0–5s
- almost all fills in 270–300s
- buys and sells with the same timestamp distribution
- zero same-market sells
- one huge clip per slot instead of staggered size
Watch the activity tab for those shifts. Style drift shows up in clocks before it shows up in a blog post.
8. What Part 5 Covers
The forensic series has now inferred:
- specialist footprint
- TWAP60 target
- signal constraints
- enter / scratch / redeem loop
Part 5 is the honest ending: what this evidence still cannot prove, what remains private, and how to verify the wallet without pretending the README is a live PnL audit.
Links
- Repo: https://github.com/abrownfox0/abrownfox001-twap60-prediction-trigger-system
- Video: https://www.youtube.com/watch?v=XzhugRL6BV4
- Live activity: https://polymarket.com/@abrownfox001?tab=activity
Not financial advice. Timestamps show process. They do not guarantee expectancy.
If you have more questions, please feel free to contact me at any time: https://t.me/abrownfox001
My Polymarket Activity: https://polymarket.com/@abrownfox001?tab=activity
#Polymarket #BTC #TWAP #TradingBot #AlgoTrading #PredictionMarkets #QuantTrading
Top comments (0)