Market Microstructure &
Mechanics
Most algorithmic failures are not caused by bad signals. They are caused by misunderstanding how markets actually function at the mechanical level.
Core Components of Mechanics
Order Formation
How participants express intent (Passive vs Aggressive). In discrete markets, Order Intent is often more informative than price itself.
Key Insight Limit vs Market orders reveal urgency.Matching & Priority
How orders interact. Systems like Price–Time Priority mean queue position often matters more than signal strength.
Failure Mode Assuming fills are instantaneous.Liquidity & Participation
Liquidity is not constant. We distinguish between Displayed, Effective, and Reactive liquidity.
Failure Mode Low participation amplifies noise and slippage.Spread & Friction
The spread is a signal, not just a cost. Spread compression often reveals instability before price does.
Key Insight Friction increases under stress.Latency & Timing
Latency is about relative timing. Signal generation vs order placement windows.
Failure Mode Designing algorithms that assume synchronous info.Venue-Specific Mechanics
Markets that look similar on the surface often behave very differently underneath. AlgoBoy treats these structural differences as distinct inputs.
Traditional Financial Markets
Key Mechanics:
- Central Limit Order Books (CLOB)
- Fragmented Venues & Routing
- Hidden Liquidity (Icebergs)
Common Pitfall: Assuming consolidated liquidity and ignoring queue competition on specific venues.
Betting Exchanges
Key Mechanics:
- Odds-based Pricing (Negotiated)
- Finite Event Horizons
- Market Closure Dynamics
Common Pitfall: Treating odds as probabilities. Liquidity collapses near the start, and late information dominates early signals.
Why Microstructure Breaks Backtests
The Silent Killer of Performance
Many backtests implicitly assume mechanics that do not exist in reality:
- Infinite Liquidity: "I can buy as much as I want at this price."
- Immediate Fills: "My order executes the millisecond the price touches."
- Stable Spreads: "The cost to trade is constant."
AlgoBoy rejects any backtest that cannot model fill uncertainty. If execution is not modeled, performance metrics are fictional.
Microstructure-Aware Design
Algorithms built with microstructure in mind tend to trade less frequently and avoid marginal signals. This results in lower headline returns but significantly higher survivability.
Studying multiple market types reveals structural truths: Betting exchanges often expose inefficiencies faster, while Financial markets hide them longer behind liquidity.
Markets are not equations. They are systems with rules, frictions, and adversarial participants.
AlgoBoy treats microstructure not as an edge, but as the ground truth upon which all edges must stand.