Realized Edge

Independent trader | Systematic research

[Realized Edge @RealizedEdge]

Realized Edge is my trading & research journal.

The idea is simple:

An edge isn't real because a backtest looks good.
It has to survive changing markets, execution, risk, and time.
Here I'll document the process:

→ Strategy research
→ Backtesting
→ Forward testing
→ Risk & portfolio management
→ What works — and what doesn't

Wins will be documented.
Failures will be documented.
Changes will have a reason.
No signals. No promises. No pretending every idea works.
The objective isn't to predict the market.
It's to identify an edge, validate it, execute it with discipline, and ultimately—
Realize it.

[Realized Edge @RealizedEdge]

REALIZED EDGE // FORWARD LOG

Baseline-01 has completed initial cloud validation.

3/3 clean qualifying unattended runs.

Tonight:
• Scheduled execution: PASS
• Duplicate decisions: 0
• Duplicate orders: 0
• Pipeline health: PASS
• Strategy changes: 0
• Real orders: 0
A prior scheduler miss remains permanently recorded. No backfill. No reconstructed data.
The strategy is frozen. Now comes the part that matters:
Let forward data decide whether the edge is real.
Research → Freeze → Forward test → Measure.

[Realized Edge @RealizedEdge]

REALIZED EDGE // FORWARD LOG

Baseline-01 just completed another clean unattended cloud run.

• Paper only
• 1 scheduled execution
• 0 duplicate decisions
• 0 duplicate orders
• Strategy unchanged
• Historical scheduler gap preserved — no backfill
• Clean cloud streak: 2/3
The goal isn’t to make the backtest look good.
It’s to prove the research process survives contact with forward data.
Research → Freeze → Forward test → Measure.
One more clean natural run completes initial cloud validation.

[Realized Edge @RealizedEdge]

REALIZED EDGE // MLB-PIT-01

First unattended point-in-time MLB market capture: PASS

• Predetermined pregame observation window
• 28 markets captured
• 9 sportsbooks
• 200 timestamped observations
• 2/2 target games met frozen eligibility rules
• Raw data hashed + immutable
• No backfill of 3 earlier missed windows
• 1 API credit consumed exactly as expected
This isn’t a betting result.
It’s infrastructure.
We’re building our own forward-only market history so future research can be tested against information that was actually available at the time — not reconstructed after the outcome.
Research → Freeze → Forward test → Measure.

[Realized Edge @RealizedEdge]

REALIZED EDGE — Forward Test Update

Baseline-01 completed its first clean, fully unattended cloud paper run tonight.

• Scheduler → execution → ledger → health checks all passed
• 0 duplicate executions
• Strategy remains frozen
• Live trading remains disabled
• Current paper equity: $4,932.86 from $5,000 inception
Cloud validation: 1/3 clean unattended runs.
No victory lap. No parameter changes.
The point is to see whether a frozen strategy survives forward data — including the ugly parts.
Research → Freeze → Forward Test → Measure → Publish.

[Realized Edge @RealizedEdge]

REALIZED EDGE — FORWARD TEST LOG

Baseline-01 infrastructure is now complete.

• Strategy logic frozen
• $5,000 simulated starting capital
• BTC / ETH / SOL
• Daily next-open execution
• Cloud paper trading automated
• Performance ledger + benchmarks automated
• Live trading disabled
First 6 forward observations:
Strategy: -5.31%
BTC B&H: -3.50%
Equal-weight BTC/ETH/SOL: -5.18%
Not impressive. And that's the point.
No changing parameters because the first week didn't cooperate.
The strategy stays frozen while the forward sample grows.
Next checkpoints: 30 / 60 / 90 days.

Build. Freeze. Measure.
Let the evidence decide.

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG — MLB02
Feature engineering complete.

• 12,144 MLB games (2021–2025)
• 3.57M Statcast pitches processed
• 178 pregame predictors
• Point-in-time / same-day leakage controls
• 2026 holdout remains sealed
• Predictive performance still unseen

Next: freeze the modeling + calibration protocol before opening 2025 validation.
Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG — SH03 CLOSED
208 oversold setups → 136 confirmations.

The event problem was solved. The edge wasn't.

Across 6 preregistered horizons, no return effect survived statistical validation. Every bootstrap 95% CI crossed zero, and all Holm-adjusted tests were non-significant.

No retuning. No cherry-picking the best horizon. No paper deployment.

SH03: NULL / NON-ACTIONABLE.

Negative results stay in the record.

Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

@DefiDragans Should be interesting to say the least

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG — A NEW ANGLE

This is getting interesting.

The project is starting to separate into 3 engines:

• Core systems → untouched forward testing
• Short-Horizon Lab → 1H crypto research
• Business → tools built from the validation process

Next experiment: can short-horizon systematic signals generate meaningful monthly returns after fees and realistic execution?

5% net/month will be a measurement benchmark — not a target we optimize toward.

If the data says 5%, we document it.
If it says 1%, we document it.
If it says negative, we document that too.

No changing parameters until the backtest looks good.
No hiding failed experiments.
No live capital.

Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG — I09 FORWARD TEST BEGINS
I09 has officially entered forward paper testing.

First unseen decision:
• BTC 20D: +21.44%
• ETH 20D: +29.03%
• SOL 20D: +33.90%
• Signal: SOL
• Execution: next daily open
• Starting capital: $5,000 paper
• Live capital: $0

Historical validation got it here. Forward data decides whether it stays.
Baseline, I06 and I07 also completed tonight’s automated runs successfully.
No tuning. No intervention. No live trading.
Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG
The research lab is built.

Baseline-01, I06, I07 and I09 are now governed by a deterministic forward-testing stack:

• paper-only execution
• normalized forward evidence
• operational audits
• hash/integrity verification
• stale-evidence guards
• 30/90/180-day checkpoints
• no automatic live approval

Historical research can generate hypotheses.
Forward data has to earn the confidence.
From here, the scarce input isn't more code.
It's time.
Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

REALIZEDEDGE | RESEARCH LOG — I09
I09's first frozen backtest is interesting — and ugly.

Daily cross-sectional relative strength across BTC / ETH / SOL:

12 bps/side:
$5,000 → $14,612 (+192%)
Sharpe: 0.68
Max drawdown: -82.1%

24 bps/side:
$5,000 → $10,708 (+114%)
The regime breakdown matters more:
2022: -73%
2023: +552%
2024: +103%
2025: +5%
2026 YTD: -22%
V2 verdict: material regime instability.

No tuning. No parameter rescue. Validation continues.
A profitable backtest isn't the same thing as a deployable strategy.

Research → Validate → Forward Test → Realize Edge

[Realized Edge @RealizedEdge]

@DefiDragans Thank you

[Realized Edge @RealizedEdge]

RealizedEdge | I06 Forward Test — Day 1
First forward-paper execution complete.

Frozen before the first trade:
• BTC / ETH / SOL
• 4H EMA20/50 trend model
• Spot long/cash only
• Independent 1/3 sleeves
• 12 bps/side modeled costs
• $5,000 starting capital

First forward signal: LONG all 3
Equity after initial entries: $4,993.85
Entry fees: $5.99
No strategy changes from here. Wins and losses both get documented.
Backtests suggest an edge. Forward data decides whether it survives.

[Realized Edge @RealizedEdge]

REALIZED EDGE // FORWARD TEST #001

Day 0. The experiment starts here.

Starting Equity: $5,000
Mode: PAPER ONLY
Live Execution: Disabled
Trades: 0
The strategy has been built and backtested.
Now I'm putting it through something a backtest can't provide:
Time.

No skipping bad trades.
No rewriting history.
No changing the rules because I don't like the results.

Every signal. Every trade. Every drawdown. Every failure.

Documented.

The objective isn't to prove the strategy works.
It's to find out if it does.
$5,000 → ?

[Realized Edge @RealizedEdge]

@moonpay Seahawks

[Realized Edge @RealizedEdge]

@unusual_whales

[Realized Edge @RealizedEdge]

@SkaterGains 2001

[Realized Edge @RealizedEdge]

@DreamilioMaker