LIVE SINCE SEPTEMBER 2026 · IBKR-DOCUMENTED TRACK RECORD

Systematic futures.
AI-optimised.
Sequentially deployed.

Nine momentum signal streams. Nikkei 225 · DAX 40 · Nasdaq 100.
One session at a time. The same capital recycled across all three markets.
No overlap. No overnight positions. Capital efficiency built in.

New to IBKR? Open an account first →
+33.49%
Return · Sep 2026
−9.91%
Max Drawdown
3.38×
Return / Max DD
62.5%
Win Rate · Daily
$12,661
NAV · Live IBKR

The Idea

Where the returns come from

Same $100,000. Three markets. 3× the capital efficiency. Nikkei, DAX and Nasdaq futures trade in three sessions that never overlap — so one pot of margin capital can back all three strategies across a trading day, instead of being split three ways and sitting idle two-thirds of the time.

$300K split three ways
Nikkei — $100K, idle 16h/day1×
DAX — $100K, idle 16h/day1×
Nasdaq — $100K, idle 16h/day1×
1×
Capital efficiency
$100K reused three times
Nikkei — $100K, then handed off3×
DAX — same $100K, then handed off3×
Nasdaq — same $100K, session closes3×
3×
Capital efficiency
3× more capital-efficient

If the three sessions are roughly independent, return scales close to 3×, volatility rises by only about 1.7× (√3), and return per unit of risk improves by about 1.7×. This is a diversification and capital-efficiency effect — not a higher-conviction signal, and not free risk.

~3×
Return, if sessions independent
~1.7×
Volatility (√3), not 3×
~1.7×
Return per unit of risk

Three sessions each take on real risk; the improvement comes from not leaving capital idle between them. Actual measured correlation across sessions is shown on the Live Account and Allocations tabs.

Human Idea, AI-Built System

The inefficiency itself was human-spotted — the same three-session intraday trend structure behind the 2003 Princeton thesis on the Research tab. AI took it from there: breaking the idea down into 9 granular signal streams, adapting and optimizing the parameters per market and session, writing the production trading code, and now running all three markets automatically, live, with no manual intervention.

Background & Research

Where the strategy comes from

Fable Fund's founder co-authored a 2003 Princeton MFin thesis, Dynamic Trend Optimization, finding statistically significant trend alpha in Nasdaq futures at a time when prevailing academic opinion leaned toward market efficiency. Later independent research has documented related intraday momentum patterns in equity index markets — cited below as supporting context, not as verification of this strategy's specific results.

2003
Princeton MFin thesis — Dynamic Trend Optimization
Written with three fellow students. Found statistically significant evidence of trend alpha in Nasdaq futures at a time when prevailing academic opinion leaned toward market efficiency.
2014–24
Independent research on intraday trend
Later academic and practitioner work documented related intraday momentum patterns in equity index markets — summarized below. None of these studies test this strategy directly; they document the broader phenomenon it relies on.
Today
Granulated and automated by AI
The same underlying trend-following logic, human-identified, was broken down by AI into 9 discrete signal streams across Nikkei, DAX and Nasdaq futures — each adapted, optimized, coded and now traded automatically, sized within a fixed daily loss budget.
Market Intraday Momentum
Gao, Han, Li & Zhou · Journal of Financial Economics, 2018
First half-hour return on the S&P 500 (SPY) predicts the last half-hour return, R² of 1.6% — also found across ten other ETFs and two major international equity index futures.
Beat the Market: An Intraday Momentum Strategy for SPY
Zarattini, Aziz & Barbon · SSRN working paper, 2024 (not peer-reviewed)
Trend-following intraday SPY strategy: 1,985% net total return (2007–early 2024), 19.6% annualized, Sharpe 1.33.
Hedging Demand and Market Intraday Momentum
Baltussen et al., 2021
Links the intraday momentum effect to dealer gamma-hedging flows — a proposed structural reason the pattern persists rather than arbitraging away.
International replication
Pacific-Basin Finance Journal, 2023
Effect present in China and Japan — not uniform across every market this strategy trades.

The Princeton Anomaly

Same capital deployed sequentially across three time zones — Asia (Nikkei) → Europe (DAX) → US (NQ). Capital recycles intraday. No overnight positions. Binding margin = worst single session only, not the sum of all three.

9
Signal Streams
3 markets × Stream 1 (daily long), Stream 2 (daily short), Stream 3 (2hr trend). 1 and 2 are mutually exclusive each day — at most one fires.
3
Sequential Sessions
Nikkei first (Tokyo), DAX second (Frankfurt), NQ last (New York). Capital freed from each closed session funds the next — no idle time.
1,340
Backtest Days
Common period Sep 2021 – Sep 2026. Start date bound by Nikkei data availability. 5 full calendar years.
0
Overnight Positions
All positions close within the session. Gap risk exists at session open only — it cannot be eliminated, only sized for with the designed stop loss.
EW 3-Market CAGR
18.9%
No leverage baseline
EW 3-Market Sharpe
2.62
Exceptional for systematic
EW Max Drawdown
−6.5%
Unlevered baseline
Smooth 30× CAGR
2,676%
Optimised, SL-constrained
Smooth Sharpe
2.69
Preserved under leverage
Smooth Max DD
−32.3%
At 30× leverage
Min Annual (Smooth)
+92.8%
Every year positive
Binding Margin
63%
NAV (Nikkei session)
Capital Efficiency

The same capital trades Nikkei at night (Tokyo), DAX at dawn (Frankfurt), Nasdaq through the day (New York) — never simultaneously. Same capital unit passes through up to 9 signals per day — 3 per market — with zero idle time between sessions.

Diversification

9 streams across 3 uncorrelated geographies + 2 strategy types (mean-reversion daily + trend-following 2hr). Portfolio vol 6.3% unlevered vs best single stream at 4.2% — but CAGR lifts from 4–10% to 18.9%.

Sequential Margin

Because sessions do not overlap, required margin is max(NQ, Nikkei, DAX) — not the sum. At $14,435 NAV with Smooth Curve, Nikkei is the binding session at $8,759 (63% of NAV), leaving $5,676 free.

AI-Optimized Execution

The inefficiency was human-spotted. AI granulated it into 9 discrete signal streams, adapted and optimized each one per market and session, wrote the production trading code, and now runs all three markets automatically, live.

Equal-Weight 3-Market Portfolio — Equity (no leverage)
Smooth Curve 30× — Equity (log scale, $15k start)
EW Drawdown
Smooth Curve Drawdown

9 Streams

Each stream run independently at unit weight (1×). Sep 2021–Sep 2026. Designed SL = mode of loss distribution (what the stop is set to). Max Lev = 7.5% daily budget ÷ designed SL.

CAGR by Stream
Stream Scorecard — Full Stats
StreamTotal RetCAGRSharpe CalmarMax DDWorst Day Des. SLMax LevWin RateTrades 202120222023202420252026
Sharpe vs CAGR
Annual Returns Heatmap

Portfolio Models

Same 9 underlying streams, three different capital deployment assumptions. Model B is the most realistic and deployable.

Model A · EW 9 Streams
+33.5%
CAGR 6.0% · Sharpe 2.61 · Max DD −2.2%
Vol 2.1% · Calmar 2.70
1/9th weight per stream, independently
Model B · EW 3 Markets ✦ Most Realistic
+137.5%
CAGR 18.9% · Sharpe 2.62 · Max DD −6.5%
Vol 6.3% · Calmar 2.92
Daily+Super combined per market, equal weight
Model C · Sequential Compound
+1,161%
CAGR 66.2% · Sharpe 2.63 · Max DD −18.3%
Vol 19.0% · Calmar 3.62
Capital compounded across all 3 markets daily
Smooth Curve 30× Optimised
2,676%
Sharpe 2.69 · Max DD −32.3% · Calmar 10.5
Worst day −7.5% designed · −12.0% historical
Min annual: +92.8% — every year positive
Equity Curves — All 3 Models (% return)
Drawdown Comparison
Model Comparison Table
ModelTotal RetCAGRSharpeMax DDCalmarAnn VolWin Rate
A · EW 9 Streams+33.5%6.0%2.61−2.2%2.702.1%51%
B · EW 3 Markets ✦+137.5%18.9%2.62−6.5%2.926.3%52%
C · Sequential+1,161%66.2%2.63−18.3%3.6219.0%51%
Smooth Curve 30× Opt+2,676%98.2%2.69−32.3%10.5——
Note on Model C: Sequential compounding means each market multiplies the day's running total. Returns are extraordinary but each market loss amplifies prior gains — the −18.3% max DD is felt on the full compounded base. Requires near-instantaneous capital transfer intraday. Model B is the most deployable.
Annual Returns — Models A, B, C
Smooth Curve vs Live Baseline (log, $15k start)

Optimal Allocations

12 distinct risk profiles. Same 9 streams, per-stream SL-based leverage bounds, 7.5% max designed daily loss budget. Select any allocation to see its equity curve, drawdown, annual returns, and per-stream risk contribution.

CAGR
—
Need 20+ days
Total Return
—
Sep 2021–Sep 2026
Sharpe
—
Daily basis
Max Drawdown
—
Peak-to-trough
Worst Day
—
Designed SL-based
Calmar
—
CAGR / Max DD
Total Leverage
—
Gross
Ann. Vol
—
Need 20+ days
Equity Curve — Log Scale —
Drawdown Profile
Per-Stream CAGR — NQ & Nikkei · DAX
Per-Stream Leverage & Max Loss Contribution
StreamWorst DayMax Solo Lev Alloc LevMax Loss% Budget Risk Bar
PORTFOLIO TOTAL — —
Risk-Return Frontier — All 12 Allocations
All 12 Allocations — Side by Side
#AllocationMax LossLeverageCAGRTotal RetMax DDSharpeCalmarVolMin Annual

Risk & Drawdown

Gap events in NQ 3 are intraday flash crashes — the position was entered at the correct signal, then the market moved 2–4% in a single 10-minute candle. The −0.25% stop filled at −3.43% due to pure slippage. Cannot be managed away — only sized for.

Smooth 30× Max DD
−32.3%
Backtest peak-to-trough
Live 46× Max DD
−54.4%
Maximum leverage — not current allocation
Days > 7.5% loss
10
Gap events in backtest
Worst streak
10 days
Aug 17 2026 (current)
Worst single day
−12.0%
Apr 7 2025 tariff shock
1 MNQ vs 2 MNQ
−14%
vs −28% on worst gap
Annual Max Drawdown by Year
Annual Return vs Intra-Year DD
Yearly Return & Drawdown
YearReturnMax DDNAV StartNAV End
NQ 3 Gap Events — Backtest Period (Sep 2021–Sep 2026) 10 events
Apr 7 2025 detail: Entry LONG at 17,842 at 18:20 · Exit at 17,229 at 18:30 — −3.43% in exactly 10 minutes. Trump Liberation Day tariff announcement. At 2 MNQ on $15k NAV: portfolio lost −28.16% in one session. At 1 MNQ: −14.08%. The only mitigation is sizing — the stop was entered correctly, slippage was unavoidable.
Risk Framework — fixed rules, every session
Hard daily loss stop
Each session has a fixed worst-day risk budget; trading stops for that session once it's hit.
Flat at session close
Positions close at the end of each session's window, removing overnight and cross-session exposure.
Volatility-scaled sizing
Position size shrinks automatically as realized volatility rises.
Independent session budgets
Nikkei, DAX and Nasdaq each have their own allocation and loss limit.

Dynamic Sizing

Your live leverage ratios (NDX 1+1+2 MNQ, NIKKEI 16+7+20 N225MC, DAX 4+2+3 FDXS at ~$15k) scaled back through time. Historical prices were lower → same margin% meant lower dollar margin per contract → more contracts per $15k. Binding margin = max session only (sequential).

MNQ margin / notional
6.9%
$4,235 / $61,570
FDXS margin / notional
5.2%
€1,349 / €25,788
N225MC margin / notional
11.1%
$461 / $4,136
Live binding margin
~110%
Of NAV (tight — sequential)
Smooth binding margin
63%
Nikkei = binding session
Contract Counts by Year — Live Leverage Ratios
YearNAV Start NQ LNQ SNQ SUP NK LNK SNK SUP DAX LDAX SDAX SUP ReturnMax DD
At current leverage ratios, $15k compounded over the 5-year backtest period would have grown to $4.0B — shown here as a leverage illustration only, not a projection or achievable outcome.
Smooth Curve — Contracts at $14,435 NAV Recommended
StreamLeverageContractsTypeNotionalSL Contrib
NQ session (US)
(1 super + 1 daily) = 2 MNQ = $8,470
Nikkei session ◄ BINDING
(12 super + 16 daily) = 28 N225MC = $12,908
DAX session (EU)
(1 super + 4 daily) = 5 FDXS = $7,722

Verification

Trade data, not marketing numbers

1
IBKR Flex Query
Trades exported directly from the broker
→
2
Structured record
Converted to a dated trade log
→
3
SHA-256 hash
Each file's fingerprint is recorded
→
4
Published record
Hash and log posted publicly
→
5
CPA reconciliation
Independent monthly check

The hash confirms a published file hasn't been altered after the fact — it does not by itself confirm the file was complete or accurate when created. That's what the broker source and CPA reconciliation are for. Account numbers and client-identifying details are never included in anything published publicly.

Important Disclosures

Past performance is not indicative of future results. Futures trading involves substantial risk of loss and is not suitable for all investors. Reusing capital across sessions increases the trading activity — and the risk exposure — run through that capital; it does not eliminate risk.

Fable Fund is not currently registered as a Commodity Trading Advisor (CTA) or Commodity Pool Operator (CPO). Any advisory or investment activity is limited to family, friends, and existing personal relationships, consistent with applicable exemptions. This site is not an offer or solicitation to any member of the public to invest, and nothing here should be construed as investment advice.

Research cited above documents general intraday momentum patterns in index and ETF markets; it does not constitute independent verification of Fable Fund's strategy or results.