Internal application
Hawk Thorne uses GEPS within its proprietary investment process. This public description covers the general concept. Operational parameters and integration with other systems remain internal.
Hawk Thorne Research / 02
GEPS
The Compounding Efficiency Frontier
Our model studies when raising a per-trade profit target delivers diminishing reductions in the number of decisions needed to reach a capital target.
PROPRIETARY MODEL · CONCEPT AND INTERNAL APPLICATION
01The research question
What does a higher trade target buy?
With constant positive returns, a higher return per decision reduces the number of decisions needed to reach a target. The benefit of each further increase diminishes.
GEPS formalises the boundary at which increasing the return by a defined increment ceases to save a specified minimum number of decisions. An extension adds a constraint on capital-path risk.
02Interactive illustration
The geometry of compounding.
Change the capital target and return per decision. The curve illustrates a property of the mathematical model, under the explicit assumptions below.
Decisions to target7.3
Decisions saved at +1 pp0.63
GEPS frontier · r*7.84%
Mathematical illustration: constant positive returns, full compounding, no losses or costs. Δr = 1 percentage point, ε = 1 decision. N(r) is a continuous decision count. The chart models neither calendar time nor actual strategy performance. r* is the boundary of the stated criterion, not a recommended profit target.
03Public definition
N(r) = ln G / ln(1 + r)
G is the ratio of target to initial capital, and r is the constant return per decision. Marginal savings are S(r) = N(r) − N(r + Δr).
The frontier r* is the largest return satisfying S(r) ≥ ε. It depends on the target G, increment Δr and threshold ε. It is not a universal optimal return.
04Application and research status
Scope of evidence
The research documentation contains the model's formalisation and a programme of simulation, market calibration and behavioural research. Operational use alone does not establish investment outperformance. This page presents neither empirical validation results nor a promise of returns.
05Limitations
The basic model assumes constant positive returns, full compounding and no costs. Actual trading includes losses, variable holding periods, costs, liquidity and changing exposure sizes. Fewer modelled decisions do not automatically mean less time or lower risk in practice.
GEPS develops the question of per-decision target selection using the mathematics of compounding. It complements research on geometric growth and exposure sizing.
Back to our approach