DealerFlow Terminal reads structure, flow, volatility, and dealer positioning across every horizon, resolves it into one calibrated read per symbol — the Diamond — and executes your rulebook at the level of automation you choose. You own the strategy and every decision; the platform is the infrastructure.
Four layers, one private deployment — from raw positioning data to a governed order, with the audit trail your operations team can stand behind.
One calibrated, directional read per symbol per horizon — a win-probability with a confidence interval and sample size, not a binary alert.
Structure, flow, volatility, technical confirmation, Bayesian calibration, and positioning — combined under weights the model tunes itself.
The Diamond Ribbon aligns every horizon from minutes to months. Cross-horizon disagreement is surfaced, never silently blended away.
Manual, semi-automated, or fully systematic — your rulebook, a firm-level risk desk, and a hard kill switch above every order.
Most emerging managers run on broker screens, spreadsheets, and instinct. These are the places it leaks — and what closes each one.
“I read price. The market is moved by positioning.”
See what dealers see — gamma, flow, and where the tape is pinned; the lens desks pay research firms for.
“Conviction is a feeling; sizing needs a number.”
A calibrated probability to size against consistently — with the honesty to say when the edge isn’t there.
“My timeframes disagree and I resolve it by gut.”
Every horizon in one view — agreement and conflict made explicit, not averaged away.
“I have edge; manual execution leaks it.”
Systematic execution of your own rules — the same entries, exits, and stops every time, no drift.
“My book sits idle between ideas.”
A systematic income overlay that works the holdings you already carry.
“Allocators want a process I can prove.”
Attribution, calibration evidence, and an audit trail — your edge, documented.
Every symbol resolves to a single object: a calibrated win-probability, tagged with its confidence interval, effective sample size, market regime, and the horizon it applies to. It is the same math for a five-minute trade and a monthly position — only the horizon changes.
The interval and effective-N tell you how much the read can be trusted. Thin evidence is labelled INSUFFICIENT DATA rather than dressed up as a number.
Reads are graded T1 / T2 / T3 by calibrated edge, so position sizing keys off conviction — and low-conviction reads are excluded, not just discouraged.
Every read is conditioned on the prevailing dealer-positioning and volatility regime — the same setup means different things in positive- versus negative-gamma tape.
Dealer positioning is one lens among six. These are the working screens of the platform, shown on your universe.
Rows are timeframes, columns are your symbols sorted by directional bias. Each tile is a Diamond — long / short probability, net bias, evidence count, and age. The pinned column is the symbol under review; disagreement across horizons is visible at a glance.
Six families each contribute an independent long and short sub-score. They combine under adaptive weights into the composite that anchors the Diamond — green ≥ 70, amber 45–69, red < 45.
The three-second verdict — magnet, flip, and the nearest walls — over a strike ladder of dealer gamma. This is Family B / F; it informs the Diamond, it is not the Diamond.
A reliability curve: when the Diamond says 60%, does it happen 60% of the time? Points on the diagonal mean the probabilities are honest. The model re-weights its families nightly to keep them there.
Weights are re-estimated nightly within governed bounds, with anomaly overrides — every change versioned in the model card.
Every symbol in your universe is scored the same way, continuously, across all horizons. Enable the families and timeframes your mandate calls for.
Market data is licensed for professional use and served inside your instance. The analytics are descriptive decision-support tools — DealerFlow Terminal does not issue recommendations.
Indicators drift; regimes rotate. Rather than trust a fixed formula, the composite measures its own hit-rate and adjusts the weight of each family every night — bounded, shrunk toward priors, and fully auditable, the way a model-risk desk would demand.
Realized outcomes feed a nightly calibration pass. Family weights move toward what is actually working — within tight, governed bounds, so nothing lurches.
Thin-evidence estimates are pulled toward conservative priors, and guardrails keep any single family from dominating or disappearing.
Every weight set, calibration curve, and anomaly override is recorded in a model card and prediction log — alongside walk-forward and backtest evidence.
Model methodology, calibration governance, and validation are detailed for diligence in our institutional overview.
The same intelligence drives three levels of automation — and you choose the level per strategy, per account. Nothing is armed that you did not configure. Above every order sits a firm-level risk desk (daily loss caps, position limits, time-slot windows) and a hard kill switch that flattens to paper instantly.
Rules live in per-ticker, versioned Trade Profiles — sizing, stops, trailing, chase and spread guards, entry offsets. One block order allocates across your accounts and SMAs at average price.
One intelligence layer, one rulebook, one risk desk — feeding three systematic engines. Each acts only on the rules you author, at the automation level you choose. It is the difference between a research feed and a desk that runs.
Reads price structure and momentum across horizons and executes qualifying setups end-to-end — entry, exit, stop, and trailing. Your discretionary edge, run with the same discipline on every name, every session.
Builds and manages multi-leg, defined-risk options structures — selecting, laddering, rolling, and reconciling as conditions move — so options exposure runs at desk scale without hand-managing every leg.
Finds eligible holdings and systematically overwrites them for yield — turning a static portfolio into a managed income stream, with timing and structure handled for you.
Rank your whole universe continuously by conviction, touch-probability, and target — and act on the handful that matter.
Firm-level caps, limits, and windows checked before an order — with a one-switch defensive posture and a hard kill.
Validate a rulebook out-of-sample on the same engine that trades it — no spreadsheet-to-live gap.
Live order, position, and P&L books, attributed per strategy, family, and profile version.
The controls a fund’s operations and diligence teams ask about first — designed in, not bolted on.
Named users authenticate through your own SSO. Access to books and accounts is granted per user, and sensitive changes — arming live, editing risk limits — require maker-checker approval.
Every order, override, arming action, and configuration change is timestamped, tamper-evident, and exportable — for your compliance file and any regulator that asks.
One fund, one instance. Dedicated compute and an isolated, encrypted database with row-level tenant enforcement — your data never shares a table with another client.
P&L attributed per book, per signal family, and per Trade Profile version — with tax-lot accounting and wash-sale detection.
Not a copy you self-host and maintain, and not a shared multi-tenant login. A dedicated instance we operate and keep current — running where you want it.
Your own compute and database. Nothing about your positions, orders, or clients touches another fund’s deployment.
Deploy inside your own AWS VPC, or in our managed environment. Your broker and market-data credentials stay yours.
We operate, monitor, and update the platform centrally — you get new families, calibration, and fixes without running a release train.
We stand up your private instance, connect your broker and data, and load your universe and mandate.
Build Trade Profiles and set the risk desk. Choose the families, timeframes, and automation level per strategy.
Run analytics live and shadow-fill your rules against real conditions — a durable “would-have-placed” record, no capital at risk.
Promote to live one account at a time, under roles, limits, and a full audit trail.
Priced per deployment, with your team’s seats and roles included. We are onboarding a small group of design partners now.
Work directly with our team to shape the roadmap. Hands-on onboarding and a direct line to the build.
A dedicated deployment for your fund, with seats and roles for your whole team.
Tell us a little about your fund and we’ll set up a scoped, 30-minute walkthrough — the Diamond, the Ribbon, your families, and the governance trail, on symbols you care about.
Thanks — we’ll reach out to schedule your private walkthrough.
No. DealerFlow Terminal is a financial-technology and analytics provider. You define and own your strategy and every investment decision; the platform executes the rules you configure at the automation level you choose. We do not issue recommendations or provide investment, legal, or tax advice.
It is the platform’s single directional read for a symbol on a given horizon — a calibrated win-probability with a confidence interval, effective sample size, market regime, and conviction tier. It is derived from six analytical families and is designed to be sized against, not blindly followed.
In your own isolated instance — dedicated compute and an encrypted database with row-level tenant enforcement, in our managed cloud or inside your own AWS VPC. Nothing is shared across clients, and your broker and market-data credentials remain yours.
Three levels, chosen per strategy and per account: manual (decision support), semi-automated (shadow / staged), and fully automated (armed). A firm-level risk desk and a hard kill switch sit above every order, arming requires maker-checker approval, and every action is audited.
Live Schwab connectivity today, with institutional FIX and multi-custodian routing — including block-and-allocate across accounts — on the near-term roadmap. Tell us your setup and we’ll map it during scoping.
Emerging managers and managed-account desks, roughly $25M–$250M AUM — large enough to need real governance and execution, but not yet running a full in-house quant and infrastructure team.
See a dedicated instance — the Diamond, the Ribbon, calibrated conviction, and governed execution — running on your universe in a 30-minute walkthrough.