1.3 Types of arbitrage Capture focuses on
Capture OS is designed around market-neutral, price-based inefficiencies. You’ll see references to the following patterns.
1.3.1 Cross-DEX arbitrage (two-leg)
The most basic type:
Same pair (e.g. SOL/USDC)
Two venues (DEX A vs DEX B)
Flow:
Buy on cheaper venue
Sell on more expensive venue
Lock in the spread (if execution is fast and size is sane)
Capture OS’s detection core is constantly scanning these pair + DEX combinations, computing net spreads after fees and slippage assumptions.
1.3.2 Triangular arbitrage (multi-leg loops)
Here you exploit a loop of three (or more) assets where the implied exchange rates are inconsistent.
Example loop:
SOL → USDC
USDC → RAY
RAY → SOL
If the product of these conversions is > 1 (after fees/slippage), there’s a loop profit.
Conceptually:
You start with 1 SOL
Go through the loop
End with 1.01 SOL
Capture OS can score these loops by chaining quotes and calculating the net return over the loop, not just on a single pair.
1.3.3 Statistical arbitrage
This is less “instant price mismatch” and more short-term mispricing relative to a model:
Assets that historically move together diverge
A pool drift vs its fair value
Temporary deviations from some equilibrium you’ve defined
Capture OS provides real-time price and spread data. A stat-arb engine can sit on top of that, using:
Correlations
Z-scores
Mean-reversion thresholds
Capture OS’s job here: feed you clean, timely market states so your own stat-arb logic isn’t blind.
1.3.4 Liquidity / yield arbitrage
This is about positioning liquidity and capital where it earns more per unit of risk, often in a market-neutral way:
Move liquidity from pool A to pool B when fees/yield per unit risk are better.
Exploit mispriced LP tokens or derivative tokens that lag their underlying.
Capture OS can help by:
Showing where volumes and spreads concentrate
Highlighting when some venue is consistently offering better realized spreads or fee capture than others.
Last updated