Three market experiments,
running on your machine.
Every agent, bidder and odds feed below is simulated locally with rule-based agents — no large language models, no token spend. Prices, revenues and arbitrage edges are all computed from real mathematical models you can inspect.
Agent-Based Market Simulator
100 heterogeneous agents · live limit order book
Fundamentalists, chartists, noise traders and market makers trade continuously on a price-time-priority LOB. Watch manias form and implode endogenously — bubble, crash and flash-crash scenarios reproduce the classic microstructure pathologies with zero LLM cost.
- ▸Candlestick chart with fundamental-value overlay
- ▸Live depth, spread, order flow and regime detection
- ▸Bubble, fundamental crash, flash crash & V-recovery
Auction Theory Lab
English · Dutch · Vickrey · First-price sealed bid
Thousands of Monte Carlo runs with private-value bidders. Verify the Revenue Equivalence Theorem, see why risk aversion favors the seller under first-price, and watch the winner's curse eat naive bidders — the reason eBay runs second-price proxy auctions.
- ▸Revenue distributions across all four formats
- ▸Risk-aversion and bidder-count controls
- ▸Live animated auction with per-bidder valuations
Sports Betting Arbitrage Scanner
Cross-bookmaker surebets & +EV value bets
Decimal odds from five sharp sportsbooks are normalized to implied probabilities and scanned for mispricing: surebets where every outcome can be backed for a guaranteed return, and positive-EV legs versus no-vig fair lines. Stake allocation included. Pure research — no real wagering.
- ▸Arb % and guaranteed ROI per event
- ▸Kelly-style stake calculator with equal-payout split
- ▸+EV table vs. no-vig consensus fair odds
Educational simulation. Market scenarios are generated by stochastic agent models; auction results are Monte Carlo estimates; odds data is synthetic but modeled on real market structure. Nothing here is investment advice and the betting module never places real wagers.