Strategy building with a verdict
Prove it before you trade it.
ConnectQ turns trading ideas into testable setups: compose them from signal blocks, prove them on months of data the setup never saw, and run only the ones that passed — on NinjaTrader, through one cloud-configured strategy.
Built for accounts that cannot afford to learn the expensive way. Why we exist →
The loop
Design. Prove. Run. In that order, every time.
Start from an idea, not a blank page
Describe the setup the way you would say it out loud — what has to be true, what fires the trade, how you get in, where the stop goes — and Forge assembles it, then reads it back to you in plain English before you keep it. Or duplicate one that already works and change a single thing.
Underneath it is the same document either way: gates, triggers, entries, brackets, exits and sizing from a catalog of 150+ market signals, nested AND/OR, no code.
Get a verdict, not a chart
Every run is split into a training window and a holdout the setup never saw. Fills are honest by construction. The output is a stamp — VALIDATED, WEAK or REJECTED — with the reasons written out.
Run only what passed
One NinjaTrader strategy, configured from the cloud. It reads the setup you built — the same blocks, the same nested gate tree — and shows its reading on the chart, bar by bar. Trading is enabled server-side, per machine and per chart slot.
Say it once, in plain words
A setup is a sentence before it is a pipeline: only look for trades when this holds, then take it when that fires, get in this way, risk it like so, get out on these. Answer that and Forge assembles the blocks for you — then reads the finished setup back in plain English, so what you meant and what it built sit side by side before you keep either. Everything is pre-filled, so you can create straight away and adjust in the Designer.
Already have one that works? Duplicate it, name the copy for whatever you are changing, and experiment there. The original is untouched, and the copy starts without a verdict — it has to earn its own.
Describe the idea
Only look for trades when…
- the 1-hour bias agrees with the tradeThe single most valuable rule in the book.
- there is no red-folder newsThe one uncontroversial filter.
- price is on my side of VWAP
- delivery has flipped my way recently
Then take the trade when…
- a fair-value gap appearsmin gap 2 refractory 10
- price crosses VWAP
- delivery flips (CISD)
What you just described
Only when the 1-hour bias agrees with the trade, and there is no red-folder news.
Take it when a fair-value gap appears.
Get in by resting a limit at the level, expiring after 40 bars.
Risk — stop 100 ticks, target 120 ticks.
Out on trail 70t · flatten after 20 bars.
Answer as much as you already know — everything is pre-filled, so you can create straight away and adjust later. Forge reads the setup back in plain English before you keep it, so what you meant and what it built are easy to compare.
- trigger
- FVG gap · min 2t · refr 10b
- entry
- LIMIT at level · expires 40b
- size
- 2 legs × 1 · tilt ×2 idle · cap 2
- exits
- SL 100t · TP 120t · trail 70t · time 20b · bias-loss
Your setup, read back by the strategy that trades it — the same gate tree, evaluated on every bar. ALL 1/1 BULLISH means the root group holds and permits longs only.
The chart reads it back to you
No black box on the right-hand side of the screen. The Bridge restates the setup it is running — the gate tree, which conditions hold, the direction they permit, and the trigger, entry, sizing and exits it will use. If it is not going to take a trade, you can see which condition is in the way.
Why it is different
Most backtesters flatter you. This one is built to disagree with you.
Honest fills, enforced
A limit order exists only after the bar that produced the signal closes. No fills on the signal bar, no profit taken on the fill bar unless the close confirms it. The three lookahead bugs that make most published backtests wrong are regression-tested out.
Holdout is not optional
You cannot run a setup without declaring which months are training and which are holdout. The verdict compares the two. A setup that only works on the data it was tuned on does not get a VALIDATED stamp.
Parity, measured rather than assumed
Switch on the diary and the Bridge writes down what it decided on every bar — each condition, each direction, the trigger and its level — and that log is compared against the engine’s, bar for bar. The last run disagreed on nothing, across five setups — four of them written to be awkward rather than flattering.
That settles interpretation, not money: fills are reconciled separately, trade by trade and to the dollar, in NinjaTrader playback, and so far that is the fair-value-gap path. The signals added most recently have a probe written and have not been through a playback yet. We would rather say which is which than let one word cover both.
Nothing arms from the chart
There is no “enable trading” checkbox on the strategy. Trading is switched on by the server for an approved machine and a specific setup, with a global kill switch. Your chart shows the full picture; the server decides what may trade.
Bars decide, ticks fill
Decisions are made on three-minute bars — the same ones on your chart. Fills are resolved on NinjaTrader’s own ticks, and before a day’s ticks may price a single trade they have to rebuild that day’s bars exactly. Ten of the first seventy failed and were thrown out. Where ambiguity still remains, every verdict lists it and resolves it against the trade, never for it.
Served, and withdrawable
A setup you deploy is served to your chart one configuration at a time and can be withdrawn in one poll. Our own managed model works the same way: your chart receives the current instructions to execute; the research, validation and rationale stay on the server.
ConnectQ Prime
A managed model, run the same way.
Prime is our own model for NQ futures, built and validated with the exact process the platform gives you — then operated for you on your NinjaTrader, on a capped number of seats. The instructions are served to your chart; the model’s research stays on the server.About Prime →
Early access
We are dogfooding. Join the list for the first seats.
Small group first, one hundred seats total at launch. Pricing is per-use — you pay for proof runs and live sessions, not a lump sum. We email only when there is something for you to try.