kernel · running · python 3.13
$ hedge --version 4.27.18 (build a8d4f1)

A quant terminal
for the
desk._

PIT fundamentals. Vectorised backtester. Deterministic seeds. FIX gateway to fourteen venues. Numbers in tabular figures. The tool a research desk would build for itself, if research desks built tools.

PIT history
35 yrs
Universe
8,427
Drift
1.8 bps
backtest · momentum_v3.ipynb
cpu · 4.2s · seed 0xbe3c
$ hedge.run(strategy='mom_v3', start='2014-01-01', end='2026-04-26')
[ok] universe · 8,427 names · PIT · survivorship-free
Sharpe (net)
1.84
CAGR (net)
14.27%
Max DD
−8.42%
Turnover
142%
[ok] tca · IS slippage 1.2 bps · backtest-to-live 1.8 bps
$ _
Used by desks at
Northwind CapitalMercury TreasuryFoundry 47 WealthPlover & Co.Field Notes CapitalTwo SigmaAQRRenaissanceCitadel SecuritiesJane StreetNorthwind CapitalMercury TreasuryFoundry 47 WealthPlover & Co.Field Notes CapitalTwo SigmaAQRRenaissanceCitadel SecuritiesJane Street

A research stack with no excuses left.

PIT, deterministic, vectorised, fitted cost model, FIX execution. The four boxes a desk has to check.

01 · Backtest

Vectorised. Deterministic. Hashed.

10-year backtest of the S&P 500 universe in 4.2 seconds. Re-runs reproduce to the cent.

Speed
4.2 s
PIT
35 yr
Survivor.
free
Drift
1.8 bp
02 · Cost model

Fitted, ADV-bucketed.

Backtest-to-live drift 1.8 bps. The reason the live PnL prints.

03 · Execution

FIX 4.4 / 5.0 to 14 venues.

Smart router. Child-order analytics. IS slippage report on every parent.

Services

Three tools. One tenant.

01

Factor research IDE

Notebook + dataframe IDE wired to point-in-time fundamentals, returns, and 14 alt-data sets.

Pricing
$840 / seat / mo
PIT · 1990-present
02

Backtester (vectorised)

Python and Julia backtester. T+0 fills, transaction-cost model, slippage by ADV bucket.

Pricing
$0.04 / cpu-hr
deterministic · seeded
03

Live execution gateway

FIX 4.4 / 5.0 SP2 to 14 venues. Smart order router, child-order analytics, IS slippage report.

Pricing
$0.0028 / share
Reg NMS · 605/606

A modern terminal. Or your internal stack.

Both run python. One reproduces to the cent.

Comparison
Hedge
Internal stack
Backtest engine
Vectorised, deterministic
For-loop, non-deterministic
Data
Point-in-time, survivorship-free
As-reported, restated
Execution
FIX 4.4 to 14 venues
Single venue, REST
Cost model
ADV-bucketed, fitted
Flat 5 bps
Reproducibility
Seeded, hash-stamped
Best-effort
From PMs & quants
"Hedge replaced our internal stack. Sharpe of the production strategy went up 0.18 because the cost model finally matched live fills."
Rafael Romero
PM, Foundry 47 Wealth
$4.27B
AUM run on hedge
1.84
avg sharpe · live
1,427
backtests / day
31 yrs
combined platform exp.

Per seat. Per cpu-hour. Per share.

No tier games. No "platform fee". The cost is the resource you used.

Personal

$840
/ seat / mo

Single-seat IDE, 5,000 cpu-hours / mo of backtest, equities universe (US + DM Europe).

  • 1 seat
  • 5,000 cpu-hr
  • US + DM Europe equities
  • PIT fundamentals
  • Email support
Request seat

Professional

most chosen
$2,840
/ seat / mo

For research desks. Adds futures, options, alt-data, and live execution gateway.

  • All Personal features
  • Futures + options
  • 14 alt-data sets
  • Live execution (FIX)
  • Slack support
Request seat

Family office

$14,200
/ mo · seats by quote

For systematic family offices and small funds. Dedicated tenant, on-prem deploy option.

  • Dedicated tenant
  • On-prem optional
  • Custom data ingest
  • Dedicated SE
  • Quarterly review
Request seat
Trust & compliance

The questions a PM should ask first.

Yes. Seeded random number generation, hashed dataset versions, and pinned library versions. Re-runs reproduce to the cent.

Request a seat.
Backtest by Friday.

A 30-minute call with a research SE. Tenant provisioned the same day. First backtest within an hour.