Evidence before confidence
Claims are separated from hypotheses, assumptions are exposed, and confidence rises only with independent evidence.
Wager Market Performance Systems
Tested on history.
Proven in markets.
WMPS Analytics is building an autonomous research and trading system for weather prediction markets. We search for repeatable advantage across historical data—then demand that it survive unseen conditions, real prices, fees, and live execution.
How we work
Scientific discipline applied to market uncertainty.
Forecast accuracy alone is not enough. WMPS connects weather evidence to the economic reality of a tradable contract, while preserving the lineage needed to challenge every conclusion.
Measure before optimizing
We profile bottlenecks, define experiments in advance, and distinguish engineering improvements from scientific results.
Preserve what was learned
Data lineage, failures, decisions, and competing interpretations remain part of the system so future work builds forward.
The system
From authoritative observations to executable decisions.
WMPS is designed as a chain of evidence. Each layer has a distinct job, and no promising model result is allowed to quietly bypass market economics or operational reality.
Observe the world
Authoritative weather records, time-causal forecasts, market histories, and preserved provenance.
Test competing beliefs
Probabilistic models, replay, calibration, ablation, and out-of-sample evaluation under frozen rules.
Prove economic value
Tradable prices, fees, liquidity, monitoring, recovery, and eventual autonomous execution.
Standard of proof
A result is not an edge until it generalizes.
Only knowable information
Replay decisions may use only evidence that was defensibly available at that historical moment.
Unseen conditions
Performance must persist beyond the dates, markets, and regimes used to design the system.
Net and executable
Commercial value is measured after fees, realistic prices, liquidity constraints, and execution assumptions.
Predeclared and auditable
Frozen experiments, explicit stop gates, and durable decision records protect the work from hindsight.
How it started
From a Texas constraint to a scientific system.
WMPS began with a practical problem. Its founder was an electrician and professional gambler looking for a way to apply years of gambling experience in Texas, where casinos are illegal.
Horse racing was the obvious alternative, but online wagering was also unavailable in Texas. By the time travel, parking, and the rest of the trip were paid for, too much of the edge was gone.
So he did what everyone else does: he asked ChatGPT. The answer did not arrive as a finished system. It arrived as a plan—and then the work began.
A question
Could gambling skill be applied to a legal, measurable market from Texas?
A single spreadsheet
The first version was small enough to see all at once: assumptions, probabilities, and results in one file.
A database
As the evidence grew, the spreadsheet gave way to structured history that could be queried, checked, and preserved.
Scripts and automation
Manual calculations became repeatable processes for collecting data, evaluating markets, and testing ideas.
WMPS today
What began as one question is now a research infrastructure built to discover whether a repeatable edge can survive live markets.
Where we stand
Development-stage by design. Honest by default.
WMPS Analytics is in active research and validation. The objective is a repeatable live trading advantage in weather markets. Until that standard is met, we describe the work as a system under test—not a proven money machine.
DomainWeather prediction markets
Research focusTemperature, precipitation & snow
Core testOut-of-sample generalization
Commercial metricNet executable profitability
End stateReliable autonomous execution
Investor conversations
Back the proof, not the pitch.
We are interested in aligned funding partners who value technical rigor, transparent evidence, and the patience to distinguish a durable system from a persuasive backtest.
Investor contactspence.wmps@gmail.com