Skip to main content
All models
H
Zirdle Research · Model HELIOS

HELIOS

5-minute intraday scalping on liquid large-caps

Sun god — always on, day-cycle, fastest light

Out-of-sample headline
+1.47%
per week at 1:5 R/R
= +3.9% total over test window
Parameters
4.4 M
Val loss
0.0868
Universe
148
Channels
12
Interval
5 min
1:1 WR
49.9%

What HELIOS does

The fastest of the Five. HELIOS predicts 30 minutes ahead on 148 liquid large-cap equities, specialising in catching short-term microstructure dislocations.

Strategy focus

Intraday momentum + microstructure dislocations

How it works

HELIOS is a factorised channel × time attention transformer. It looks at a 200-bar context window of 12 input channels, and predicts the full distribution of the next 6 bars.

Input
(B, 200, 12)
OHLCV + indicators
Patch embed
per-channel linear
patch_len=10 → (B, 20, C, D)
Factorized attention ×N
space → time → FFN
channel × time attention
Pool
last patch, mean
over channel axis
Quantile head
(B, 6, 7)
q05 · q50 · q95

Forward pass: input OHLCV + indicator channels → patch-embedded → alternating space/time attention layers → pooled to a single embedding → MLP head predicts 7 quantile levels per horizon step.

Input channels

Every bar is z-score normalised per-window using context-only statistics, so the model sees relative moves rather than absolute prices.

closeopenhighlowvolumeRSI-14MACDMACD-histStoch-KATR-14Bollinger %BOBV

Output

Quantile levels per horizon step0.05 · 0.10 · 0.25 · 0.50 · 0.75 · 0.90 · 0.95
Horizon6 steps
LossPinball (quantile)
Directional callsign of (q50 − entry)

Data & training

Universe

148 symbols

TOP_500_LIQUID

Symbols are the liquid, tradeable instruments that form HELIOS's training population. Each prediction cycle runs forward-tests on the same universe so performance numbers aren't cherry-picked.

Why this training window

We deliberately exclude pre-2010 data. Pre-decimalisation (fractions until April 2001), pre-HFT regime (≤2007), and the 2007-2009 credit crisis reflect a market structure that no longer exists. At 4.4 M parameters, burning capacity on that regime is noise competing for weights with the current market — López de Prado (2018) identifies regime non-stationarity as the primary failure mode of financial ML.

Sensitivity: a pilot trained on 2003-2022 ran validation loss 10-12% worse than the 2010-onwards run.

Live performance

Every prediction HELIOS makes is logged and barrier-evaluated in real time. These numbers reflect closed trades only and update continuously.

Past performance does not guarantee future results. Forward-test outcomes reflect 1:3 R/R barrier simulation; live outcomes may differ due to slippage, spreads, and liquidity.

Honest limitations

Bull-regime test exposure

The 15-week out-of-sample window was a bull-biased period. Model performance in bear regimes (2022-Q4 style) is unverified and likely materially lower.

Transaction costs not modelled

Headline returns exclude round-trip fees (~0.05-0.30% on these instruments). Real-world returns will be lower; headline numbers are upper bounds.

Same-symbol train/test

The model was trained and tested on the same symbol universe (time-split). We validate temporal generalisation, not universe generalisation to unseen tickers.

What we're confident about

Directional skill above random chance (symmetric 1:1 R/R win rate is 49.9%) and reproducible training: we've retrained from scratch 3+ times with identical hyperparameters within ±1% val loss.

Other models

Zirdle - Global Credit Market Research Platform

Financial data, research tools and educational resources for global credit markets

All financial data and research tools are provided for informational and educational purposes only. Nothing on this site constitutes financial advice or a recommendation to buy, sell, or hold any security.

Zirdle Ltd | Company No. 16806866 | Registered in England & Wales | 1st Floor, 124 Cleveland Street, London, W1T 6PG

© 2026 Zirdle. All rights reserved.