Universal Financial Markets Analysis

Mark the setup, estimate the odds, then decide whether to trade

MarketPawns marks market structure with one method, Universal Market Geometry (UMG). Neural networks look at that marked setup and estimate how likely the next step is — a reversal, the next target, or the setup running out. Those numbers go into an expected-value check. If the check is positive, a trading agent can send the order to a broker. Analysts work on the same charts, so research and execution stay in one place.

See the platform Project overview

How MarketPawns works

One sequence: mark the structure, estimate the odds, then size and send the trade.

Universal Market Geometry (UMG)

UMG is the method for reading a chart. It marks structure, turning points, and targets with the same visual rules on any instrument or timeframe. Because the rules stay the same, two models can be compared and discussed without starting from scratch.

How this is used

Probability estimates

Neural networks do not press buy or sell. They read a UMG model and estimate the chance of specific outcomes: a reversal, reaching the next target, or the setup running out. Those numbers are what the expected-value check uses. Up to 20 networks can work on one strategy, each on its own question. They are retrained when new strategies and data are added.

How a trade is prepared

Shared research

Traders look at the same UMG markings, so a discussion is about the same model. Draft ideas go to IdeaBox, other analysts comment, and work is ranked with ELO.

Community tools

Pattern recognition

The scanner looks for UMG models across instruments and marks them the same way every time. You get a list of setups instead of searching the same charts by hand.

Where this runs

Trading agents

A trading agent is a program that runs a strategy on a broker account: it sends the order, keeps the risk limits on it, and watches the position. When a setup passes the expected-value check, it can go to an agent. The trader still owns the strategy. The agent follows it, and execution is watched live.

How orders go out

Risk controls

For each trade the system calculates position size from the strategy rules, so a move to the invalidation level stays inside the allowed loss.

Project status

Services

Each tool covers one part of the work: learning, analysis, testing, discussion, or monitoring.

MarketPawns Platform

The main workspace: charts with UMG models, probability scores, and signals on several asset classes.

See the platform

Starter Academy

A course that teaches UMG from first charts to working analysis, before you use the rest of the platform.

Open the course

Sandbox

Try signals and strategies on live charts without putting money in. Forex, stocks, crypto, and metals.

Open Sandbox

IdeaBox

Draft trading ideas. Other analysts can comment before an idea becomes a strategy.

Open IdeaBox

Community Board

The forum for UMG work: notes, model discussion, and ELO-ranked analysts.

Open the board

Knowledge Wiki

The reference for UMG rules, platform functions, and trading procedures.

Open the wiki

Trading Monitor

Trading agents, fills, and account state across brokers, on one screen.

See project status

How a trade is prepared and executed

From the marked chart to an order, then again at the next target.

First the chart is marked with a UMG model: structure, invalidation, and targets. Neural networks then answer specific questions about that model — for example, whether price is likely to reach the next target. Those answers go into an expected-value check. If the check is positive, a trading agent can send the order. When price later reaches a target, the same questions are asked again, and the trade is kept or closed.

Neural networks

Neural networks look at a marked UMG model and answer specific questions: will the move reverse, will price reach the first target, should the trade be closed. Each answer is a probability, not an order.

A separate expected-value check then uses that probability together with the distance to the target and the stop. Only if the result is positive can the trade go out. When price reaches the next target, the questions are asked again.

  • Recheck At each target
  • Entry rule Expected value > 0
  • Networks used Up to 20

Data

The networks are trained on real UMG models. Each model is stored as geometry, ratios, and hundreds of measured numbers, so two setups can be compared on the same terms.

That archive is what the networks learn from. When a new setup appears, they score it as a UMG model, using the same measured numbers.

  • Parameters per model 300+
  • Historical models 10 million+
  • UMG types 7

Trading agents

A trading agent is the program that talks to the broker and carries out the strategy. If the expected-value check is positive, the agent can send the order. Size and risk limits are already on the order.

Each agent can run one strategy, or several, without mixing them. It then watches the position: fills, size, and whether the trade is still worth holding.

  • Execution Broker-connected agents
  • Position size From strategy rules
  • Connection Broker APIs
  • Monitoring Live

Integrations

The same analysis core can be used on the main platform or in a partner setup. APIs and broker links sit around that core.

Each deployment uses the same UMG.

  • API REST
  • White-label Available
  • Partners Brokers and platforms
  • Hosting Cloud or dedicated
Project status

Investment and partnership

MarketPawns has been building UMG for more than ten years. That method now runs in a live system: the chart is marked, the setup is scored, and a trading agent can send the trade.

  • The research is in one platform and used in daily work.
  • Broker links, live trading agents, and an ELO-ranked analyst group are already running.
  • Around that core are learning, testing, and community tools.

PAWN-COIN

A public sale of PAWN-COIN is planned. It is not open now. In the white paper, PAWN-COIN is the unit of the platform economy and a future public-blockchain asset linked to how the platform is used.

Read White Paper Read Lite Paper Token Readiness PAWN-COIN

Platform snapshot

Current figures.

10+
Years of work
Up to 20
Networks on one strategy
10M+
Analyzed models
99.5%
System uptime
300+
Parameters per model
20+
Trading strategies

What is already in place

The method, live trading, and partner connections are already running. The analysis core does not need to be rebuilt.

UMG

The analysis method. More than ten years of use and revision.

Live operations

Trading agents run on real accounts. Results are recorded.

Architecture

Neural scoring runs on more than one server, with failover. Partners can connect through APIs.