Paper: arXiv 2610.01115
Authors: Nicolò Bonacorsi
Abstract
In this paper we study whether a trading signal can accumulate enough statistical evidence for reliable deployment before its economic value decays. We solve the problem by measuring statistical evidence and remaining opportunity on a common Kullback-Leibler information scale, obtaining an exact feasibility threshold in a canonical Gaussian model and general information lower bounds beyond it. We show that finite information budgets generate survival frontiers, search penalties, and a market equilibrium in which arbitrage activity shortens the lifetime of certifiable opportunities; a retrospective funding-rate study illustrates the role of persistence.
Complexity vs Empirical Score
- Math Complexity: 9.0/10
- Empirical Rigor: 6.0/10
- Quadrant: Holy Grail — high math complexity, high empirical rigor
Why this score: This paper presents a highly mathematical framework for evaluating alpha capacity, integrating information theory and stochastic processes. While it includes empirical illustrations, the core contribution is theoretical, making it strong in both math and rigor. The concept of ‘Certified Alpha Capacity’ and its implications for market equilibrium are novel.
Research Flowchart
flowchart TD
A["Research Goal: Certified Alpha Capacity & Arbitrage under Decay"] --> B{"Key Methodology: K-L Information Scale"};
B --> C["Data/Inputs: Trading Signal, Market Data"];
C --> D{"Computational Process: Feasibility Threshold & Info Lower Bounds"};
D --> E["Key Findings: Survival Frontiers, Search Penalties, Market Equilibrium"];
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes Certified Alpha Capacity (CAC), a framework that compares statistical evidence generated by a trading signal to its remaining economic lifetime using KL information.
Source sentence
We introduce Certified Alpha Capacity (CAC) to compare these two quantities.
Assumes 6
- Deterministic information lifetime A for the certification/deployment rule. model
Source sentence
For deterministic information lifetime A, a certification/deployment rule is a pair (S, δ), where S ∈TA and δ ∈{0, 1} is GS-measurable.
- Near the economic endpoint the drift gap follows ∆t = c(T −t)^p + o((T −t)^p) with p>0 and Σt converges to a positive‑definite matrix. model
Source sentence
Assume that as t ↑T, for some p > 0 and nonzero c ∈Rd, ∆t = c(T −t)^p + o((T −t)^p), Σt →ΣT ≻0.
- Instantaneous trading objective is quadratic: qᵀµ_t − ½ qᵀΓ_t q with Γ_t ≻0. model
Source sentence
Let the instantaneous trading objective be q⊤µt −1/2 q⊤Γt q, Γt ≻0,
- The economic boundary model b = (b_t) defines KL divergence relative to b. model
Source sentence
Throughout this section, a superscript (b) on an information quantity means that KL divergence is measured relative to the economic‑boundary model b = (b_t).
Show 2 more
- Funding‑rate events require a continuous 180‑day pre‑event history and a seven‑day outcome window. data
Source sentence
An event requires a continuous 180‑day pre‑event history, current absolute funding at or above its trailing 95th percentile, a seven‑day cooldown, and a complete seven‑day outcome window.
- Implementation costs (turnover, short‑sale costs, capacity, price impact) vary substantially across anomaly strategies. execution
Source sentence
The implementation coefficient matters empirically because turnover, short‑sale costs, capacity, and price impact vary substantially across anomaly strategies.
Data
- Open Source Asset Pricing data set alternative public
Source sentence
Using the Open Source Asset Pricing data set [1], let top_i denote the originally reported in‑sample t‑statistic for predictor i...
- BTC‑USDT and ETH‑USDT perpetual‑futures funding histories other public
Source sentence
The data contain BTC‑USDT and ETH‑USDT perpetual‑futures funding histories from Binance, Bybit, Gate, HTX, KuCoin, and MEXC, observed at eight‑hour settlements.
Costs
Not applicable
Finds 5
- Exact Gaussian certification threshold: deployment probability πα(A)=Φ(√A−z_{1−α}) and feasibility requires A ≥ A_crit(α,β) = (z_{1−α}+z_{1−β})²/2. theoretical
Source sentence
... the largest possible probability of deployment by information time A among all level‑α rules is πα(A)=Φ(√A−z_{1−α}). Consequently ... A
- Survival frontier: for instantaneous Sharpe S=2 and M=100 candidates, the minimum half‑life required is 7.245 years. theoretical
Source sentence
... The highlighted point is S = 2, M = 100, for which the minimum half‑life is 7.245 years.
- Empirical association: Spearman correlation between the funding‑rate lifetime score and next‑seven‑day signed funding is 0.476 (BTC/ETH combined). statistical
Source sentence
Spearman association with next‑seven‑day signed funding
- Equilibrium insight: below the certification frontier reliable post‑certification value is zero; above it rents finance entry and crowding reduces remaining information budget. theoretical
Source sentence
Below the frontier, reliable post‑certification value is zero. Above the frontier, rents finance entry and crowding reduces the remaining information budget.
- Remaining KL information near death scales as I_rem(t) ≈ cᵀΣ_T⁻¹c /[2(2p+1)] (T−t)^{2p+1}; for p=1, I_rem(t) ∼ cᵀΣ_T⁻¹c·(T−t)³/6. theoretical
Source sentence
/[2(2p+1)] (T−t)^{2p+1} + o((T−t)^{2p+1}). In particular, for a transverse crossing p = 1, I(b)_rem(t) ∼
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 25 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.