YOUR FIRST QUANT DESK
Understand the strategy.
Monetize the next idea.
Bring your own rules or start with a candidate. Make one deliberate change, then compare the evidence with simply holding SPY.
YOUR ACCOUNT
Welcome to your Quant Desk
Sign in to keep your strategies, test history and AI conversations together.
Verify your email
Need help or backtest access? Contact us on WhatsApp.
Have an invited account token?
ACCOUNT SECURITY
Change your password
Choose a password of at least 12 characters.
ADMIN
Account AI allowances
Set each account’s daily AI tokens and responses. Set either limit to zero to pause its AI access. Usage already recorded is retained; audit credits are separate.
01 / START
Where are you starting?
STRATEGY RULES · GUIDED CHECKLIST
Tell me how your strategy works.
Describe it in your own words. Then complete the checklist below so we can prepare precise rules for you to review.
Use Ask your Quant Desk below for conversation and AI drafting when Gemini is available, or complete this checklist directly. This form does not send your description to an LLM. Only confirmed structured rules are tested.
Example 1 · SPY + bond allocation — prefilled for you to edit.
What information should I include?
- Instrument: what do you trade? This pilot supports SPY and the BND bond ETF.
- Entry and exit: exact conditions, indicator type and lookback.
- Position size: percentage in SPY and BND, or cash for a trend strategy.
- Timing: next-open execution, or a fixed allocation rebalancing schedule.
- Evaluation dates and starting capital.
- Financing: choose 1× without borrowing, or up to 3× with an explicit manual borrowing APR. State any extra stops, shorting or combined signals; these are not supported by the current engine.
Let’s make the rules explicit.
Your description is context. Select the supported rule set that fully represents it. If your strategy needs other indicators, stop losses, shorting or assets beyond SPY/BND, keep it as a research idea; this pilot cannot test it faithfully yet.
TA test: 5 audit credits per audit or bounded search. Draft preparation and saving rules are free. A failed calculation refunds all 5 reserved credits.
Defaults: SPY benchmark includes dividends; bond portfolios include BND price changes and distributions; residual cash earns zero; next-open trades; 1 bp commission + 1 bp half-spread + 1 bp slippage per trade. No taxes or market impact. Review all values before confirming.
Advanced: import a QI JSON specification
Use a supported QI specification if you already have one.
02 / UNDERSTAND & FREEZE
Strategy rules
Signals use completed closes; trades execute next open. Cash earns zero. Each trade includes 1 bp commission, 1 bp half-spread and 1 bp slippage by default. Dividend entitlements and payment dates are included. No orders are placed.
This TA test uses 5 audit credits. A bounded TA search also uses 5 credits total. Saving rules is free; calculation failures refund the full reserved amount.
Exact saved specification
03 / ONE GUIDED CHANGE
Test a hypothesis, keep the original.
PARAMETER RESEARCH
Find and validate parameters
Save original rules above, then compare a bounded set of settings. Selection uses earlier data; later periods show historical walk-forward diagnostics. Every trial is saved. No untouched holdout or guaranteed SPY outperformance.
Trend lookbacks: 20, 50, 75, 100, 150, 200, 252. SPY allocations: 0%, 25%, 50%, 60%, 75%, 90%, 100%; the remainder is BND for bond portfolios. Your saved value is also included. Other rules remain fixed. Requires about seven years plus warm-up. Standard searches use 1 audit credit; RSI and momentum-breakout searches use 5 credits total, disclosed before running.
Parameter research results
Calmar is CAGR divided by absolute maximum drawdown for the same period. It is unavailable when the annualization window is too short, drawdown is zero, or equity becomes insolvent; hover an unavailable metric for its reason.
Later-period validation
Each period starts independently with cash. Parameters are selected using only the preceding 1,261 sessions. These are historical diagnostics on previously inspected data, not a continuous portfolio or a fresh holdout.
Full-window development exploration
Every grid member is shown below. This table does not select the walk-forward parameters. Green historical numbers are not a deployment recommendation.
All trials, training selections, neighbors, cost checks and assumptions
04 / EVIDENCE
Calculating your Reality Check
04 / EVIDENCE
Your Reality Check
Equal starting capital. Session spacing; labels are exchange dates. Development research, not an untouched holdout.
Assumptions and unavailable metrics
Did this clarify your next step?
YOUR EVIDENCE HISTORY
Test history
Saved strategy versions
PILOT OFFER · PRICE HYPOTHESIS
Reality Check · US$99
Standard tests use 1 credit; RSI and momentum-breakout tests use 5 credits per audit or search. Your report includes the SPY benchmark and a next-test plan.
YOUR QUANT DESK ASSISTANT
Ask your Quant Desk
Chat here to ask a question, clarify a rule, explore an idea or understand your results. General questions are welcome too.
Strategy draft
Reviewing a draft does not save rules or run a test. Save the original only after checking the complete specification.
AI details & usage
Gemini is the selected provider for chat inside QI. Free-tier availability and limits depend on the linked Google project; QI’s allowance does not change Google’s quota or billing settings.
Messages and any context you choose to include are sent to Google Gemini using QI’s server-side account. Gemini free-tier content may be used to improve Google’s products. Conversations are saved in your QI account.
AI chat is enabled by default. Sending a question or preparing a draft shares this conversation and your selected context with Google Gemini. Free-tier content may be used to improve Google’s products. Uncheck Enable AI chat to opt out.
The assistant has no live web or market-price access. Its answers can be mistaken. Only QI’s confirmed, supported rules are backtested; chatting does not place orders or spend audit credits.
Context is optional. Without it, the assistant receives this conversation and QI’s supported capabilities, rather than your saved strategies or reports.
Each answer or draft request uses your AI allowance. Backtest audit credits are separate.