Glossary
Technical and financial terms used across Mechaniq · 65 entries
This page explains the jargon you'll find on the dashboard, in alerts, and in backtest reports. Terms link to each other where relevant. If a term you're looking for is missing, let us know.
#
0DTE
trading
Zero days to expiration — an option that expires the same trading day. The S026 setup sells 0DTE SPX iron condors: the whole trade opens and closes within one session, so there is no overnight gap risk, but end-of-day gamma is extreme.
A
ATR
technical
Average True Range — measures average daily price range over N days (typically 14 or 20). Used to size stop-losses and detect volatility compression. Low ATR ratio (short-vs-long) often precedes breakouts.
Attribution
analysis
Statistical test that measures whether each individual metric (Tight, Uptrend, RS, etc.) actually predicts forward return — using Spearman ρ between metric score and 5/10/30/60-day returns. Heavily-weighted metrics with grey ρ are candidates for re-weighting.
B
Backtest
analysis
Running the signal logic against historical price data to estimate how it would have performed. Results are indicative, not predictive — regime changes can break past patterns.
Breakout
signal
Price moves above a defined resistance level (the pivot) on strong volume (≥1.5× 50-day average). Highest-conviction tier — entry zone for swing trades.
C
Capitulation
market
A sharp, high-volume sell-off where holders give up at once, often marking a short-term extreme. The put-selling setups (S011/S012) sell premium into capitulation within an uptrend, betting the panic is overdone.
Catalyst
metric
Score component reflecting recent fundamental news (especially EPS growth ≥25% YoY). Stocks scoring high on Catalyst often coincide with earnings beats or guidance upgrades.
Composite score
metric
Weighted average of all metric scores (0-100). Determines tier assignment: ≥70 watchlist, ≥78 imminent, breakout requires extra pivot+volume conditions.
CVaR
metric
Conditional Value at Risk (a.k.a. Expected Shortfall) — the average outcome of the worst 5% of sessions, not just the threshold. It measures the tail directly, which is what matters for left-skewed premium selling where the average and win-rate flatter the strategy. Needs a large sample to mean anything.
D
Defined risk
trading
A position whose worst case is known and capped in advance — here by the long wings. Max risk = (wing width − credit) × multiplier; the structure can never lose more than that, however far the underlying moves.
Delta
technical
An option's sensitivity to a $1 move in the underlying, also read as a rough probability of finishing in-the-money. A 0.16-delta short strike sits further out-of-the-money (safer, less premium) than a 0.25-delta one (closer, more premium, more risk).
Distribution day
technical
Down-day where volume is above the 50-day average. >3 distribution days in 25 sessions triggers a veto (institutional selling pressure).
E
EMA
technical
Exponential Moving Average — weighted moving average that reacts faster to recent prices than a simple MA. EMA-8/21/65 stack ('ribbon') is a trend-quality signal.
Expectancy
metric
Average profit or loss per trade over many trades, net of frictions. A binding gate: a setup must show positive expectancy, not merely a high win-rate, to advance.
F
Failed
method
A setup whose pre-registered trial did not clear a gate. The result is logged in the open, not buried — most ideas fail, by design, and that transparent fail record is the credibility of the lab.
Feature-gating
system
Server-side decorator (@require_subscription) that blocks 13 paid routes for users without an active or trialing Stripe subscription. Falls back to a free-tier upgrade banner on /breakout instead of a 403.
Fill model
trading
The assumed execution price when simulating a trade. Mechaniq stores each trade at the mid (optimistic) and at the full bid/ask cross (worst) and headlines a realistic case between them, so the spread's impact is bracketed and nothing is flattered.
Finnhub
data
Free financial-data API used to source the US Common Stocks universe (~4900 tickers). Backup for fundamentals when yfinance fails.
Forward return
analysis
Price change from signal date to N trading days later. Used to measure how often a signal is followed by profit — and how big.
Friction
trading
Combined cost of slippage, commission, and spread. Backtests subtract 0.5% from every trade return to simulate realistic execution.
Frozen gate
method
A pass/fail criterion locked at pre-registration and never edited afterward. If a setup fails, the gate stays as written and the postmortem is public; recalibration means a new setup ID, not a quiet rewrite.
G
Gamma
technical
How fast an option's delta changes as the underlying moves. Near expiry gamma is enormous, so a small adverse move can swing a 0DTE position to its max loss quickly — the structural danger time-stops address.
H
Hit rate
analysis
Percentage of signals that produced a return above a threshold (typically +10% within N days). Higher = more reliable signal.
Hold-to-settlement
trading
Letting a position run to the exchange's cash-settlement print at expiry rather than closing early. S026's benchmark cells hold to settlement so the managed cells can be measured against the tail you avoid by closing early.
I
Imminent
signal
Tier between Watchlist and Breakout. Composite ≥78 AND price within 2% below the pivot. Used as 'prepare order' signal.
Implied volatility (IV)
metric
The market's forward-looking estimate of how much an underlying will move, backed out of option prices; higher IV means richer premium. Mechaniq flags where a result leans on an IV proxy rather than true historical IV.
Iron condor
trading
A defined-risk options structure: sell an out-of-the-money put spread and an out-of-the-money call spread at once. It profits if the underlying stays in a range and both short strikes expire worthless; the long 'wings' cap the loss if it doesn't.
L
Liquidity filter
data
Universe-level filter requiring ≥$5M average daily dollar-volume (20-day). Excludes thin-traded stocks where breakouts often fail due to slippage.
M
Max drawdown
metric
The largest peak-to-trough drop in cumulative P&L — how much a strategy gave back from a prior high, and a direct read on the worst losing run an investor would have lived through.
Mechaniq
method
This service. Mechanical (rules-based) signal engine for US stocks. 'Mechaniq' (with a q) = the discipline of letting a defined system make calls instead of discretionary opinion. No tips, no hype, no black-box weights — every signal is reproducible from the published rules.
Multi-tenant
engineering
Every alert, backtest, and signal-outcome row carries a user_id foreign key. Two customers running on the same server see only their own history; no data crosses tenants. Implemented via the user_id column on alert_log, signal_outcome, and backtest_run.
O
OHLCV
data
Open, High, Low, Close, Volume — standard daily bar format. Source for all technical metrics.
OOS / Out-of-sample
analysis
Backtest period that wasn't used for parameter tuning. A signal that looks great in-sample but fails OOS is overfit.
Overfitting
analysis
Tuning a strategy so tightly to past data that it captures noise instead of signal. Detected via OOS-tests and walk-forward.
P
Pivot
technical
Resistance level identified from prior swing highs (60-day window excluding current bar). Breakout = close > pivot on strong volume.
Pre-registration
method
Writing down a setup's exact rules and pass/fail gate BEFORE any test runs, then freezing them. It removes the freedom to rationalise a result after the fact — the discipline the whole lab is built on.
Profit-take
trading
Closing a position early once a set fraction of the credit has been captured (e.g. 25% or 50%) instead of holding to expiry. For premium selling it trades some average gain for a smaller tail — a core variable the S026 matrix tests.
Put credit spread
trading
Sell a put and buy a further-out-of-the-money put for protection — a defined-risk way to collect premium that profits if the underlying stays above the short strike.
Q
Quick-screen
method
Stage 1 of the pipeline: a first gate on historical data. Most ideas end here, by design — shown as 'Failed' on the progress rail.
R
Regime
market
Broad market environment classified by SPY's position vs its SMA200. 'Up' = SPY > SMA200 (trending), 'Down' = SPY < SMA200 (defensive). Signals behave differently per regime.
Regime gate
system
Configurable rule that suppresses Imminent/Breakout alerts when SPY is in down-regime. Watchlist signals continue as early radar.
Regime gate (Mechaniq)
system
Mechaniq's specific implementation of the regime concept: when SPY < SMA200, Imminent/Breakout alerts are suppressed but Watchlist continues. Avoids whipsaws in bear markets while keeping the early-radar function. Toggle in breakout_config.json.
RS / Relative Strength
metric
Stock's price performance vs the universe over 3/6/12 months (weighted 50/30/20%). Higher = leadership stock per O'Neill.
Russell 1000 / 2000
universe
Index of largest 1000 (or next 2000) US stocks. Originally our universe source; replaced with Finnhub after iShares API was blocked.
S
Setup
method
Mechaniq's term for a candidate strategy (e.g. S011, S026). Each setup is pre-registered with a frozen gate, then run through the pipeline stages; the public record shows where every setup stands.
Slippage
trading
Difference between expected and actual execution price. Higher in low-volume stocks and during volatile market open/close.
SMA
technical
Simple Moving Average. SMA200 = 200-day average close, the canonical trend filter. Stock below SMA200 triggers the below-sma200 veto.
Spearman ρ (rho)
analysis
Rank-correlation coefficient (-1 to +1). Used in attribution to test whether a metric's score actually predicts forward return. |ρ|<0.10 at N<200 is statistically weak.
SPX / SPXW
market
S&P 500 index options. SPXW are the daily/weekly-expiry, European, cash-settled contracts S026 trades — no early assignment, no single-stock risk, and no shares ever change hands.
SPY
benchmark
S&P 500 ETF — proxy for the broad US market. Its SMA200 trend defines the regime; its returns are subtracted from individual stocks to compute Relative Strength.
Sweep
analysis
Running the same backtest multiple times across a parameter range (e.g. imminent_min ∈ [70,75,80]) to see how counts and returns change.
T
Tail risk
analysis
The risk concentrated in rare, severe outcomes. Premium-selling payoffs are left-skewed (many small wins, occasional large losses), so Mechaniq's gates are tail-aware — judged on CVaR and worst-session severity, never win-rate or arithmetic mean alone.
Theta
technical
The rate at which an option loses value as time passes, all else equal. Premium-selling strategies are 'long theta' — they profit from this decay, which is fastest near expiry (the reason 0DTE exists).
Tier
signal
Classification of a signal by strength: Watchlist (composite ≥70), Imminent (≥78 + near pivot), Breakout (>pivot + volume confirmation).
Tight structure
metric
Score component for volatility compression — short ATR (10) vs long ATR (50) ratio below 0.9 for N consecutive days. Tight structure often precedes range expansion.
Tiingo
data
Power tier vendor for EOD prices (full US universe, survivorship-bias-free) and News API. Source for the Breakout strategy and the lab/ pipeline's quickkill experiments. Fundamentals add-on is separate (not yet active).
Time-stop
trading
Closing a position at a fixed clock time rather than a price level. S026 closes its 0DTE condors by 15:30 ET to avoid the violent end-of-day gamma near expiry — the genuine 0DTE risk control, used instead of a price stop.
Trade simulation
method
Stage 3 of the pipeline: running a setup forward on live data with simulated money before any real capital. S026 and the put-selling setups are in trade simulation now; results accumulate publicly toward the gate.
V
Vectorization
engineering
Computing rolling metrics (SMA/EMA/ATR/RS) once per symbol as full time-series, then looking up O(1) per scoring event. ~4× speedup vs recomputing per event.
Veto
system
Hard rule that forces composite to 0 regardless of metric scores. Examples: below_sma200, >3 distribution days, earnings_within_5_days.
W
Walk-forward
analysis
Split data into chronological folds; tune parameters on fold N (train), measure performance on fold N+1 (test). Reveals whether tuning generalizes or just overfits the training period.
Watchlist
signal
Lowest tier (composite ≥70). Early-radar signal — stock has woken up but isn't yet ready to enter. Stays active across all market regimes.
Wing
trading
The protective long option in a spread, bought further out-of-the-money than the short strike. The wing width (e.g. $10 or $20 in index points) sets the maximum loss — wider wings cost more but cap a smaller share of capital.
Y
yfinance
data
Open-source library wrapping Yahoo Finance — primary OHLCV source for all 5000+ symbols. Fundamentals come from yfinance with Finnhub fallback.
Mechaniq provides information, not investment advice. We do not execute trades. Past results are no guarantee for future performance. You are solely responsible for your trading decisions.
Lab · Methodology · Greeks Lab · About · FAQ · Glossary · Privacy · Terms · mechaniq.trade © 2026
Updated: 07 Jul 2026, 07:05