SIGNALDELTA
AI Trading System v1.0
PAPER TRADING SYSTEM ACTIVE
─ AN AUTOMATED TRADING PLATFORM

Two AI systems.
Same market.
The delta is the proof.

Real money. Verified results. Every week. Connect your exchange — your funds never leave it.

LANE 1 · LANE 2 · Δ
Δ > 0
REPRESENTATIVE · BUILD PHASE
─ ARCHITECTURE
2 lanes
Parallel AI systems · same trade infrastructure
─ LANE 1 · TARGET
60%
Industry baseline 50% · learning loop refining toward target
─ CADENCE
Weekly Δ
Verified results published every Sunday
─ FIXED UNIVERSE · CRYPTO
3 pairs
BTC · ETH · SOL on Alpaca
─ FIXED UNIVERSE · EQUITIES
Watchlist
Large-cap + high-volatility growth · activated at account funding floor
─ CONCENTRATION
Intentional
Narrow universe gives the learning loop enough volume per asset to find real patterns rather than noise

Most trading bots run one strategy. SignalDelta runs two — at the same time, on the same market.

A baseline AI system runs alongside a second, independent system on the same trade infrastructure. The gap between their returns — once Lane 2 activates and is measured across hundreds of trades — is empirical evidence the second layer adds alpha.

LANE 1 — BASELINE · LIVE
A research-backed signal stack, refining toward target.

A multi-indicator AI signal stack built on industry-standard components — RSI, EMA, VWAP, MACD. Industry baseline win rate is ~50%. The 60% target is drawn from published peer-reviewed research on this indicator combination — Mengshetti et al. (2024), Journal of Open Innovation 10(3), Article 100334 — which reported 60.63% profitable trades and a 1.882 profit factor at the composite level. Universe and methodology differ from SignalDelta's; the figure informs design, not forecast. A learning loop reviews every week's trades and refines the rule set toward that target as the system accumulates history. The engine evaluates the market every 60 seconds and tracks up to 126 data points per cycle across crypto, stocks, news, market context, capital flow, and engine output.

Every trade carries 34 properties — every indicator reading, market condition, news event, stop management decision, and outcome metric. News is AI-classified twice daily into seven categories with impact ratings. Sharpe ratios are computed three ways. Returns are time-weighted and adjusted for capital flows.

Every Sunday, the system reviews the week's trades and refines its own rule set. The audit trail is complete.

Every trade, every rule, every config change is a permanent record — the system cross-references the full history to find patterns no human would spot.

Cadence
Evaluates every 60 seconds
Tracked
Up to 126 data points per cycle
Per Trade
34 properties, no nulls
News
AI-classified twice daily · 7 categories
Review
Weekly self-refinement · Sunday 18:00 UTC
LANE 2 — PREDICTIVE LAYERShadow Mode
A second AI layer, gated until it earns its place.

A second AI system runs in parallel to Lane 1 on the same trade infrastructure. Both contribute to the final trading decision once Lane 2 is cleared for live deployment.

The mechanism is withheld pending provisional patent. What's public is the discipline around it.

Lane 2 is built but not live. It runs in shadow during the build phase — the system records and evaluates its output without acting on it. It will only activate against real capital after it clears a statistical-power threshold most retail systems never test against.

No vanity sample sizes. No early activation. No exceptions.

Status
Built · shadow mode · activation gated
Criterion
Must demonstrate measurable lift over Lane 1
Sample
Properly powered · enforced before go-live
When live
Lane 2 contribution to returns = the delta
A live map of system activity shows what the engine is tracking in real time. Visible in the simulation portal ↗
Δ — MEASURED OUTCOME
When Lane 2 activates, the delta — Lane 2 return minus Lane 1 return — will be published weekly across every live trade. A persistent positive delta is empirical proof the second layer works. Until then, both lanes run in shadow and the math has to clear the gate.
Δ > 0

Numbers, not adjectives.

Every trade is logged. Every week is published. Every figure is verifiable. The system is currently in paper mode during the build phase — once deployed, results are live and on the record.

BUILD PHASE PAPER SYSTEM ACTIVE
PHASE
1.1 — Engine Build Complete · 48-Hour Stability Test Passed
MODE
Paper Trading
CAPITAL DEPLOYED
None — gated on performance criteria
CUSTODY MODEL
Trade-only API keys · funds never leave your exchange
PERFORMANCE DROPS
Sunday 18:00 UTC · Lane 1, Lane 2, Δ
OPERATOR
Brian Albert
─ TWO WAYS IN

Try the demo. Watch the channel.

A live, public, multi-month track record — accumulating in front of you. Either the delta exists, or it doesn't.

View live simulation portal