PelosiAlpha

PelosiTracker

Auditable research, paper tracking, and controlled US execution
CONTROLLED POLYMARKET US EXECUTION
Loading thresholds…

Polymarket US research & execution

The established engine remains unchanged. A separate, auditable execution policy maps its signals to exact US/mobile contracts and enforces your bankroll limits.
POLYMARKET US
DeploymentLoading…
API keyLoading…
TradingLoading…
Safety boundary: live entries and live cash-outs are disarmed after every restart and when the selected 30-minute to 4-hour latch expires. Dry-run is the default. Orders are bounded fill-or-kill limits, every live order is previewed, and the emergency stop touches only orders created here.
Set up trading Pick simulated or real money, connect an account, set your allocation and a risk style.
Start setup

Get started with trading

Collects the minimum needed to run: a mode, an account key for live orders, an allocation ceiling, and a risk style. Everything set here can be changed afterwards in the Policy view, and nothing is saved until the final step.
DRY RUN
How should this lane trade?

Trading performance

Running execution results for positions managed by this execution service. W–L–P counts closed trades by realized profit, loss, or push; open trades affect marked net but do not enter the record until they close.
No managed trade results yet.
Total net = realized net + current executable cash-out value minus entry cost on priced open positions. The tally buttons start a new display session only — for reading one night's dry-run profit on its own — and never delete positions, journal evidence, the datasheet, or daily-loss safeguards.

Result charts

Drawn from the same retained closed trades as the datasheet, using each trade's recorded entry settings. Closed trades only — open positions carry no realized net yet. A short run of trades is weak evidence regardless of how the line slopes.
Charts draw once retained trade rows have loaded.
Cumulative realized net by lane
Realized net by line type
Entry edge vs realized net

Trade performance datasheet

Every retained managed trade, its execution result, line type, entry-time signal, and immutable settings snapshot. A win means positive realized after-cost P/L; it is not a claim about the final game result.
Open this tab or refresh to load the retained trade ledger.
Settings producing the filtered results
Groups use the exact policy snapshot recorded at entry. Read win rate beside sample size, net, and after-cost ROI; small groups are weak evidence.
Opened Mode Result Event / selection Line Buy Stake Net Edge Quality Settings
No trade data loaded.

Reactive settings advisor

Compares realized after-cost managed trades across whole events, estimates how many logged opportunities would clear alternative execution filters, and reports the fitted sport model's shadow readiness. It does not alter probability, edge, signal quality, calibration, or engine gates.
Line types to compare
Reactive, not automatic: changing the goal or requested frequency generates a new auditable comparison. All-data mode pools live and simulated outcomes for candidate discovery but cannot one-click validate a live policy. A recommendation changes nothing until you explicitly preview or apply it.
Run an analysis to compare the current policy with available trade and model evidence.
No recommendation has been generated in this page session.
Exports the complete scope, diagnostics, candidate frontier, and validation evidence for independent review.

Automatic trade controls

These are execution limits around the existing model—not new probability, edge, quality, or calibration formulas.
DISARMED
Loading execution policy…
Effective authorization by line and game stage
Resolved by the server from the global fallback switch, the global line-type list, and every line/stage profile. This is what will actually execute.
Tap a lane to edit it. Dry-run and live automation are independent and can run together.
Dry-run quick control Start or stop the selected lane without changing its other saved settings.
Switch to the live lane to arm or disarm real orders. Dry-run automation can run without arming.
Recommended rollout: start with strict mode or the core-gates preset and run dry-run until the journal shows exact event/line mappings and exits you agree with. Selective mode changes only which existing engine gate results this execution sidecar requires; it never changes their calculations. Live mode still requires positive source edge and every authenticated US execution safeguard.
Execution policy Six steps: lane & capital, authorization, bankroll, exit behaviour, advanced gates, review.
Edit execution policy

Execution policy setup

These are execution limits around the existing model—not new probability, edge, quality, or calibration formulas. Each step saves to the lane selected in the auto trader.
Back to auto trader
Dry-run lane Loading capital… Loading…

Managed positions

Managed trades show exact entry quantity and cost. The authenticated account snapshot below also shows phone/manual positions read-only; those are never adopted for automatic trading.
Clear trades Dry run is one atomic reset: it switches automation off and force-clears simulations without consulting a market. Live submits real sell attempts. Use Reset live tally above only after every live position is closed.
Action target
Polymarket US account positions Not synchronized yet.
Sync to load the authoritative account portfolio.
No managed positions.

Execution journal

A running window into candidates, gate rejections, mapping decisions, previews, fills, marks, and cash-outs. Rejections include the observed value and configured threshold so you can see exactly which check failed.
Loading…
No automation decisions yet.

Sport Model Lab

Local research only. It records the existing engine output beside sport-specific live state and price movement, then labels observations after the final. The MLB candidate is state-aware, not a momentum score. It never changes a production probability, activates a model, or authorizes a trade.
No silent learning: fitting creates a versioned comparison candidate with a chronological event-block test. Promotion remains a separate reviewed process and is not available from this screen.
Synchronize local and hosted research evidence
This transfer works in both directions. To create the central hosted dataset, export once from the workstation and merge it on the website. Dry runs started later from a phone are then retained in the hosted dry-run lane. To bring those newer hosted sessions back to this computer, export from the website and merge the archive locally.
  1. Export hereDownloads closed trades, entry settings, audit evidence, adaptive labels, and Model Lab rows from both lanes.
  2. Open the other installationWebsite for local-to-hosted; workstation for hosted-to-local.
  3. Merge onceChecksums and source-lane IDs make repeat imports safe. Live and dry-run evidence stay separate.
Loading sport readiness and observations...

Collected segments

No observations yet.

Manual candidate fit

No candidate has been fitted in this session.

Recent live-state observations

Waiting for moneyline observations.

US/mobile sports inventory

Raw Polymarket US long-side bid/ask and listed outcomes. The execution layer may map established signals to these contracts; it never feeds these prices back into or changes the calculation engine.

Bot Activity

Loading...

Historical Odds