A quantitative model · out-of-sample backtest

What the model did on data it had never seen.

A backtest of XQE’s validated factor ensemble on held-out data it never trained on (2023–2025), net of costs. Live results are expected lower — we show the full picture, not a highlight reel.

1.0×1.1×1.2×1.3×202320242025
Out-of-sample backtest of the validated-factor ensemble — a quintile long-short reference strategy, net of 18 bps. Backtest Sharpe ≈ 1.8; live / paper target 0.8–1.1. Simulated, not real-money, not financial advice.
simulated · out-of-sample backtest
+0.0%
Net return / yr
simulated · out-of-sample backtest
0.0%
Maximum drawdown
forward expectation
0.8–1.1
Sharpe ratio — live target
What XQE is

A quantitative model for US mid-cap equities — it scores the market across dozens of factors every day. It shows you what it sees; you decide.

~400 stocksdozens of factorsdaily refresh
≈400 stocks, scored every day · simulated
How it reads the market

Factors in. A clear, ranked view out.

01

Reads the factors

Each stock is measured across dozens of fundamental, technical and behavioural factors.

02

Scores the whole market

Factors combine into a single 0–100 read for every stock, ranked side by side.

03

Surfaces conviction

The names where signal is strongest rise to the top — with the reasoning visible.

momentum — weak → strongvalue — pricier → cheaperstrong & cheap — focus today
lower scorehigher scoredot size = conviction
today’s cross-section · simulated · each dot is one stock · names shown in the demo
Why you can trust it

Most ideas are rejected. Only a few survive every gate.

A factor only reaches the model after passing a gauntlet of statistical gates. Each stage throws candidates out — what’s left has earned its place.

0
proposed
0
beats null + significant (t>3)
0
holds out-of-sample
0
robust (DSR, FDR, sign)
0
real signal (beta, de-dup)
0
survives costs → live
illustrative of the funnel · numbers vary by campaign
null-feature testt-stat > 324-fold walk-forwardtrain/test holdoutDeflated Sharpe > 0.95yearly sign-consistencyFDR / Benjamini–Hochbergbeta-residualdegeneracy + de-dupnet-of-cost economicslive health + reconciliation
What you’ll get

Explore the model’s view yourself.

Soon you’ll open any stock and see its score, the factors behind it, and where it sits in the market — the same view the model uses. Nothing hidden.

● demo opening soon

Model output · one stockpreview · illustrative
78/ 100 composite
Value+0.82
Quality+0.71
Momentum+0.64
Revisions+0.58
Low vol+0.46
Still running

The track record keeps growing.

XQE is live in paper trading right now. New results come in every day — and we’ll keep updating them in the open. Follow AI Cyber Media to watch it unfold.

live · updating
About

AI Cyber Media is a technology company building rigorous, transparent AI systems. XQE is our quantitative research model for public equities — built on point-in-time data, validated out-of-sample, and shown with its reasoning in plain view.

Visit aicybermedia.com.au ↗