Skip to content

CGQI

Menu

Technology

How the research is built.

A model is easier to evaluate when its inputs, assumptions and test conditions are visible. Follow those choices from a market observation to a result.

Home   /   Technology

Methodology

Three stages of a quantitative study

Market data and timing

Keep each observation’s venue, quote currency, observation time and release time. Choose a consistent interval before comparing equities and digital assets. A late release or revised series can change what was knowable at the decision point.

Algorithm and liquidity research

Specify the reference price before comparing executions. A record of the order size, fill quantities, available liquidity and explicit charges lets a researcher examine why two apparently similar orders produced different costs.

Factor design and testing

Document what a factor measures and when its inputs become available. Fix the rule, define a later evaluation period and rerun the test under different cost assumptions. Keep a record of each change so the result is reproducible.

What to put in a research record

KeepWhy it matters
Source and observation timeMakes the information set at the decision date traceable.
Universe and exclusionsShows which assets were eligible and which were left out.
Rule and evaluation windowSeparates the period used to design the rule from the period used to evaluate it.
Benchmark and costsShows how much of the outcome depends on the method of implementation.

See the method in practice.

Three short notes work through a market clock, a quoted spread and a reporting lag.

Browse perspectives