01 What This Platform Does
Auris is a quantitative signal generation platform. It collects price, volume, fundamental, macro, and sentiment data for any publicly traded stock, then runs that data through a multi-factor scoring model to produce a signal — a probabilistic estimate of whether a stock is likely to outperform its peers over the next 5 trading days.
Think of it as a systematic research assistant that applies the same analytical framework a quantitative hedge fund uses, running it consistently across every stock in your universe without emotional bias.
Important: This is a short-term momentum and factor signal, not a long-term investment recommendation. It does not predict earnings, FDA approvals, acquisitions, or any fundamental event. It answers one specific question: given current price behavior, volume, fundamentals, and macro conditions, does the next 5 days look better or worse for this stock versus its peers?
02 How to Read the Signal Levels
Strong Buy
72–100
Multiple factors aligned bullishly. Highest probability of near-term outperformance.
Buy
58–71
Most signals positive. Reasonable lean bullish with moderate conviction.
Neutral
42–57
Mixed or flat signals. No meaningful directional edge. Monitor rather than act.
Sell
28–41
Momentum weakening, technical deterioration. Lean toward avoiding or reducing.
Strong Sell
0–27
Multiple factors aligned bearishly. Highest probability of near-term underperformance.
Score above 70 — strong momentum, broad factor agreement, volume confirming
70–100
Score 55–70 — most signals bullish, minor contradictions present
55–69
Score 40–55 — mixed signals, no clear directional advantage
40–54
Score 30–40 — deteriorating momentum, risk signals elevated
30–39
Score below 30 — strong bearish alignment across multiple factor categories
0–29
03 Timeframe & Intended Use
| Signal Type | Primary Horizon | Typical Use Case | Rebalance Frequency |
|---|
| Strong Buy / Strong Sell | 3–7 trading days | Short-term positioning, swing trades, tactical hedges | Weekly or when signal reverses |
| Buy / Sell | 5–15 trading days | Entry/exit timing for existing theses | Bi-weekly |
| Neutral | N/A | Hold existing position; do not initiate | Re-evaluate weekly |
| Sector sweep (all 20+) | 5–10 trading days | Identify relative value within a sector | Weekly scan recommended |
04 Reading the Table Columns
5D Mom / 20D Mom
The percentage price change over the last 5 and 20 trading days respectively. Positive values mean the stock has been rising. The model weighs 20D momentum more heavily than 1D because it captures a more durable trend rather than single-day noise.
RSI (Relative Strength Index)
A 0–100 oscillator measuring the speed and magnitude of recent price moves. RSI above 70 is traditionally "overbought" — but in a strong momentum regime, the model treats RSI 60–75 as a confirmation of strength, not a reversal signal. RSI below 30 means heavily sold off and may indicate recovery potential. RSI between 40–60 is neutral territory.
Rel Vol (Relative Volume)
Today's trading volume divided by the 20-day average volume. A reading of 2.0× means twice the normal amount of shares are trading hands. Volume spikes on up days confirm institutional conviction. Volume spikes on down days confirm distribution (selling by large players). Look for Rel Vol above 1.5× when a signal fires — it means the move has participation behind it.
EPS Surp (Earnings Surprise)
The percentage by which the company's most recent earnings beat or missed analyst consensus estimates. A reading of +15% means earnings came in 15% above what analysts expected. Post-earnings announcement drift — the tendency for stocks to continue moving in the direction of their earnings surprise for 30–60 days — is one of the most well-documented anomalies in finance. This is a meaningful input.
Short% (Short Interest)
The percentage of the float that is currently sold short. High short interest (above 15%) combined with rising price creates short squeeze risk — as shorts are forced to buy to cover losses, it accelerates the upward move. Low short interest means less of a squeeze dynamic but also that the stock has less skepticism priced in.
ML Score
The composite score from 0–100. This is the weighted combination of all 45 inputs, run through a logistic function. A score of 50 represents no directional edge. Every point above 50 represents incrementally stronger bullish alignment; every point below 50 represents incrementally stronger bearish alignment. The model is nonlinear — moving from 65 to 75 is more significant than moving from 50 to 60.
05 How Macro Factors Affect Scores
The macro environment module applies a regime adjustment to all scores. Here is how each macro input affects the model:
- 10-Year Treasury Yield rising rapidly — reduces scores for high-growth, high-multiple stocks (MedTech, biotech, tech) because rising rates compress valuations. Increases scores for financials and energy.
- VIX (Volatility Index) above 25 — activates a defensive regime. All scores are pulled toward neutral. Strong Buy signals require even higher conviction to maintain their rating under high VIX.
- Oil price trend — rising oil is a tailwind for energy names and a headwind for airlines, logistics, and consumer discretionary.
- USD strength — a strengthening dollar typically hurts multinational revenue for large-cap US exporters, but benefits domestic-focused companies.
- Credit spreads widening — a leading indicator of financial stress. When high-yield spreads widen significantly, the model applies a broad caution adjustment.
Macro data is sourced live from Finnhub via ETF proxies (TLT for rates, GLD for gold, USO for oil). All data is current as of the last market close with real-time intraday quotes.
06 The SIM Badge
If a stock card or table row shows an amber SIM badge, it means Finnhub could not return data for that ticker — usually because the symbol is not covered (e.g. some ADRs, pink sheets, or very small caps). The ML scoring still runs using a deterministic synthetic generator, but signals will not reflect real market conditions. This is rare for major US-listed stocks.
07 What the Model Does Not Know
- Whether a company is fundamentally overvalued or undervalued (P/E, P/B, EV/EBITDA are not used)
- Specific upcoming FDA approval dates, drug trial readouts, or litigation outcomes
- Management quality, competitive dynamics, or industry disruption risk
- Events occurring after the data was last fetched (a scandal breaking after market close will not be reflected until the next scan)
- International political risk, sanctions, or currency crises
For a complete investment process, combine this signal with fundamental research, valuation analysis, and risk management discipline. The signal tells you when to look at a stock; fundamental analysis tells you whether to own it.