01 — Quant team

Quantitative research at the Ghosal Fund.

Research and realtime trading. The team designs the systems that run both.

RESEARCH NOTES · 3
RESEARCH SYSTEMS · BUILT
REALTIME TRADING ENGINE

02 — What the team does

Research, research systems, and the realtime trading engine.

Each strand has a defined job: test a strategy idea, run that test at scale, or connect live data to orders, risk, and monitoring.

01
Research
Take a strategy from question to backtest or paper-trading result, then present the numbers and limits to the team.
02
Research systems
Built
Runs historical and walk-forward tests on the same code paths strategies use before capital is allocated.
03
Realtime trading engine
Connects live market data to strategy logic, order routing, risk checks, and monitoring.

03 — Realtime trading engine

How live trading is wired

The realtime engine ingests market data, runs approved strategies, routes orders with risk limits, and feeds monitoring so the desk can see fills and exposure.

01
Market data
02
Strategy
03
Orders and risk
04
Monitoring

04 — Research

Published strategy notes

Three notes document the question, method, and printed backtest or paper-trading figures for crypto futures, shipping changepoints, and equity momentum.

01 — Crypto Futures

Crypto Futures

A composite technical index on top-15 crypto perpetual futures, tested with 180-day walk-forward backtests and separate paper trading on TradingView.

Walk-forward backtest (costs not included): P&L −$5.70, return −0.7125%, Sharpe −0.3225, average drawdown 0.45%.

Paper trading (documented separately from the backtest): starting balance $100,000, balance $125,423.83 (+25.4%), maximum drawdown 6.1%.

KAMA
VWAP
MFI
Williams VIX Fix
ADX
Rolling z-scores
Bayesian optimiser
Weighted linear model
Sigmoid
Index [0, 1]

Backtest · Paper trading

Return, backtest against paper trading

−0.7125%−0.7125%

Walk-forward backtest

Backtest

+25.4%+25.4%

TradingView paper account

Paper trading

% · Walk-forward, 180-day window, about July 2025 to January 2026. Paper-trading dates are not stated.Crypto Futures research note. Back testing metrics and forward testing metrics. Separate tests.

Return, backtest against paper trading. Backtest · Paper trading. Walk-forward, 180-day window, about July 2025 to January 2026. Paper-trading dates are not stated.. Units: %. Source: Crypto Futures research note. Back testing metrics and forward testing metrics. Separate tests.
Series	Kind	Value
Walk-forward backtest	Backtest	−0.7125%
TradingView paper account	Paper trading	+25.4%

Backtest · Paper trading

Drawdown, by the measure each test prints

0.45%0.45%

Average drawdown

Backtest

6.1%6.1%

Maximum drawdown

Paper trading

% · Same tests as the return chart. The measures are not the same.Crypto Futures research note. Average drawdown on the backtest. Maximum drawdown on the paper account.

Drawdown, by the measure each test prints. Backtest · Paper trading. Same tests as the return chart. The measures are not the same.. Units: %. Source: Crypto Futures research note. Average drawdown on the backtest. Maximum drawdown on the paper account.
Series	Kind	Value
Average drawdown	Backtest	0.45%
Maximum drawdown	Paper trading	6.1%

Paper trading

Paper account, start and end

The two balances printed for the TradingView paper account. This is not a path.

Start

100,000

End

125,423.83

USD · Paper trading on TradingViewCrypto Futures, forward testing metrics. P&L +25,423.83, return +25.4%, maximum drawdown 6.1%.

Paper trading account balances. Start 100,000 USD. End 125,423.83 USD. Source: Crypto Futures, forward testing metrics.
Start	100,000
End	125,423.83

Backtest: −$5.70; Backtest profit and loss
Backtest: −0.3225; Backtest Sharpe
Paper trading: +$25,423.83; Paper-trading profit and loss

Read the note

02 — Changepoints

Changepoints

5 October 2026

Two related tests on VLCC tanker AIS and energy benchmarks: FOCuS changepoints on USO, and a Mahalanobis stall signal on listed tanker names against BWET.

Strategy 1 (USO): 43 trades, 62.8% win rate, 65.23% cumulative return over two years. MCMC summary: 99.1% probability of profit, 1.69× expected multiplier, 5% floor 1.15×.

Strategy 2 (stalled tankers vs BWET): 26.30% net return, 61.54% win rate, 0.94% annual alpha versus BWET with the 94-hour AIS lag applied. No second MCMC block is reported for this leg.

AIS every 30 min
Grouped by day
FOCuS
Page-CUSUM
O(log n)
Mahalanobis
Z > 1.8

Persian Gulf to Singapore

01
Ras Tanura
02
Strait of Hormuz
03
Malacca Strait
04
Singapore

Waypoints named in the Changepoints note. Not a map.

Two-year simulation · Backtest

Return over two years, two separate tests

65.23%65.23%

Strategy 1, USO

Two-year simulation

26.30%26.30%

Strategy 2, stalled ships

Backtest

% · Two years of historical dataChangepoints research note. 5 October 2026. Strategy 1 is a simulation. Strategy 2 is labelled a backtest.

Return over two years, two separate tests. Two-year simulation · Backtest. Two years of historical data. Units: %. Source: Changepoints research note. 5 October 2026. Strategy 1 is a simulation. Strategy 2 is labelled a backtest.
Series	Kind	Value
Strategy 1, USO	Two-year simulation	65.23%
Strategy 2, stalled ships	Backtest	26.30%

Win rate: 62.8%; Strategy 1, 43 trades
Backtest: 61.54%; Strategy 2 win rate
Backtest: 0.94%; Strategy 2 annualized alpha

MCMC check · Strategy 1

Three printed figures

Probability of profit: 99.1%
Expected multiplier: 1.69×
5% risk floor: 1.15×

Strategy 1, two-year simulation. The note gives no distribution, so none is drawn. Source: Changepoints, Strategy 1 results.

Read the note

03 — Momentum

Momentum

A quarterly sector-then-stock momentum book (up to 50 names, costs not included) with a drawdown-based exposure scaler, tested against SPY over the full sample and on rolling one-year windows.

Full period (costs not included): strategy 66.3% versus S&P 500 40.7% (+25.6% relative), across 16 rebalances.

Rolling one-year windows (36 windows from 2022-01-01 to 2024-12-01): mean strategy return +9.32% versus SPY +9.65%, alpha −0.33%; best window +47.63% (1 May 2023–30 Apr 2024); worst −26.33% (1 Jun 2022–31 May 2023); positive windows 25 of 36.

Sector score
Top 5 sectors
Stock momentum
70 / 30 score
Up to 50 stocks
Quarterly rebalance

Scores

The two formulas

Sector score     0.25·R(3M) + 0.25·R(6M) + 0.50·R(12M)
Stock momentum   0.25·R(3M) + 0.5·R(6M) + 0.25·R(12M)
Score            70% momentum + 30% technical stability

The top 5 sectors become the stock universe. Up to 50 stocks, weighted by score, rebalanced quarterly.

Full period

Full period, portfolio against the S&P 500

66.3%66.3%

Portfolio

Full period

40.7%40.7%

S&P 500

Full period

% · From 2021. Costs off. 16 rebalances.Momentum research note. Full-period results. Outperformance +25.6%.

Full period, portfolio against the S&P 500. Full period. From 2021. Costs off. 16 rebalances.. Units: %. Source: Momentum research note. Full-period results. Outperformance +25.6%.
Series	Kind	Value
Portfolio	Full period	66.3%
S&P 500	Full period	40.7%

Full period: +25.6%; Outperformance
Full period: 16; Rebalances
Full period: Off; Costs

Rolling one-year windows

36 windows, summarised

The note prints the mean, the best, the worst, and the count of positive windows. It does not print the 36 returns, so they are not plotted.

Mean portfolio: +9.32%+9.32%
Mean SPY: +9.65%+9.65%
Mean alpha: −0.33%−0.33%
Best: +47.63%+47.63%
Worst: −26.33%−26.33%
Positive windows: 25 of 3625 of 36

Best window: 1 May 2023 to 30 April 2024. Worst window: 1 June 2022 to 31 May 2023.

Rolling one-year windows from 2022-01-01 to 2024-12-01 (36 windows). Percent returns.

Rolling one-year windows, 2022-01-01 to 2024-12-01. Mean portfolio +9.32%. Mean SPY +9.65%. Mean alpha −0.33%. Best +47.63%. Worst −26.33%. Positive in 25 of 36.
Mean portfolio	+9.32%
Mean SPY	+9.65%
Mean alpha	−0.33%
Best	+47.63%
Worst	−26.33%
Positive windows	25 of 36

Rule

Drawdown steps

Documented exposure cuts at 10% and 15% drawdown; exposure resets at the next quarterly rebalance. Rule table only — not a plotted equity curve.

01
10% drawdown
Exposure 0.5
02
15% drawdown
Exposure 0.25
03
Next rebalance
Resets

Source: Momentum pitch, risk management. At 10% drawdown, exposure 0.5. At 15%, exposure 0.25. Reset at the next rebalance.

Read the note

05 — Where members go

After the team.

Industry

J.P. Morgan
Nethermind
BlackRock
bp
Microsoft
Palantir
STOR-iLancaster University
STOR-i

Masters

Imperial College London
London School of Economics
University of Warwick

Member destinations. Not affiliated with or endorsed by these organisations.

06 — Join

Open roles on the quant desk

Two paths: build the shared systems or research and test new strategies.

Quant Developer

Designs the systems for research and realtime trading.

Apply

Quant Researcher

Designs and tests strategies, including the futures, shipping, and momentum work.

Apply

enquiries@ghosal.fund

Quantitative research at the Ghosal Fund.

Research, research systems, and the realtime trading engine.

Research

Research systems

Realtime trading engine

How live trading is wired

Published strategy notes

Crypto Futures

Return, backtest against paper trading

Drawdown, by the measure each test prints

Paper account, start and end

Changepoints

Return over two years, two separate tests

Three printed figures

Momentum

The two formulas

Full period, portfolio against the S&P 500

36 windows, summarised

Drawdown steps

After the team.

Industry

Masters

Open roles on the quant desk

Quant Developer

Quant Researcher