Skip to content

LEVELS.md

Latest commit

 

History

History
204 lines (160 loc) · 11.6 KB

LEVELS.md

File metadata and controls

204 lines (160 loc) · 11.6 KB
Raw

Price levels

Audience: User

Find prices the market may care about — formula pivots, retested support/resistance, multi-method confluence, and volume-profile value areas. All of these tools are read-only and exploratory: they map structure, they do not issue trade signals.

Dense terms: Pivot points · Support and resistance · Confluence · Volume profile (POC/VAH/VAL)

Related: Sample trade · CLI · Glossary

Choosing a tool

Tool Source of levels Use when you want…
pivot_compute_points Formula pivots from the last completed OHLC bar Classic floor-trader PP/R/S targets for the session
support_resistance_levels Historical retests and reactions Data-driven levels the market actually respected
confluence_levels Pivots + S/R + Fibonacci + volume profile combined High-probability zones where several methods agree
volume_profile_levels Traded volume distribution by price POC / value area to frame fair-value and acceptance

All four accept --detail compact|standard|full. Compact returns the nearest, most actionable levels; --detail full returns the raw diagnostic payload.

pivot_compute_points

Computes pivot levels from the last completed bar on timeframe.

mtdata-cli pivot_compute_points EURUSD --timeframe D1 --json
mtdata-cli pivot_compute_points EURUSD --timeframe D1 --method camarilla --json
mtdata-cli pivot_compute_points EURUSD --timeframe D1 --end 2026-07-03T20:00:00Z --json
  • timeframe defaults to D1 because daily pivots are the common floor-trader convention.
  • end / as_of is an optional historical cutoff, consistent with confluence_levels. The source bar is the last completed bar whose close is at or before that instant; later bars are not used. The payload reports both the requested cutoff and the effective source-bar close.
  • method selects one of classic, fibonacci, camarilla, woodie, demark. Omit it to return every method at --detail standard/full; --detail compact returns the classic pivots only.
  • DeMark pivots depend on whether the bar closed above, below, or at its open. Its R1/S1 formulas are canonical; the returned PP=X/4 is a common retail platform extension and is labeled pivot_convention=retail_x_over_4_extension.

Use support_resistance_levels for complementary data-driven levels.

support_resistance_levels

Detects support/resistance levels around the current price from historical structure.

mtdata-cli support_resistance_levels EURUSD --timeframe H1 --lookback 200 --json
mtdata-cli support_resistance_levels EURUSD --timeframe auto --detail standard --json
Parameter Default Description
timeframe H1 Bar timeframe, or auto to merge levels from M15, H1, H4, and D1.
lookback 200 Maximum bars used to detect levels (after any start/end window).
start / end — Optional time window (e.g. "1 month ago", "now").
tolerance_pct 0.15 Price-cluster tolerance in percentage points (0.15 = 0.15%) when grouping touches into one level.
min_touches 2 Minimum tests before a price qualifies as a level.
max_levels 4 Maximum levels to return.
max_distance_pct 5.0 Keep levels within this % of current price; pass none/null (CLI) or None (Python) for all.
volume_weighting off Set auto to weight touches by traded volume.
reaction_bars 6 Bars after a test used to measure bounce strength.
adx_period 14 ADX period for pre-test trend-strength weighting.
decay_half_life_bars — Optional exponential time decay so recent tests matter more.

Score combines repeated tests of a level, bounce strength after each test (normalized by ATR), pre-test ADX trend strength, exponential time decay, and ATR-filtered Fibonacci retracement/extension levels from the most relevant swing.

Output: ranked supports and resistances lists. Each row's price is value, with score, a type reflecting current price geometry (support below price, resistance above), and a dominant_source showing whether the historical tests mostly behaved as support or resistance. --detail standard adds Fibonacci swing levels; --detail full returns the raw diagnostic payload. Compact output also reports lookback_bars.

Use confluence_levels to cluster these levels with pivots and Fibonacci into scored consensus zones.

confluence_levels

Finds nearby high-confluence zones where multiple level methods agree. It combines formula pivots, touch-derived support/resistance, Fibonacci swing levels, and optional volume-profile levels.

mtdata-cli confluence_levels EURUSD --pivot-timeframe D1 --sr-timeframe auto --json
mtdata-cli confluence_levels EURUSD --min-source-families 2 --max-levels 5 --json
Parameter Default Description
pivot_timeframe D1 Timeframe for the formula pivot inputs.
sr_timeframe auto Timeframe for support/resistance inputs (auto merges M15–D1).
lookback 200 Bars of history for the S/R component.
tolerance_pct 0.15 Clustering tolerance in percentage points (0.15 = 0.15%) for merging levels into a zone.
tolerance_points — Broker-point count; overrides tolerance_pct. On five-digit EURUSD, 10 × point(0.00001) = 0.0001 price width.
min_touches 2 Minimum touches for the S/R component.
max_levels 5 Maximum confluence zones to return.
max_distance_pct 5.0 Keep zones within this % of current price; pass none/null (CLI) or None (Python) for all.
min_source_families 2 Require this many independent families per zone (use 1 to include single-family clusters).
pivot_method — Restrict pivots to one method (classic, fibonacci, …).
volume_weighting off Set auto to volume-weight the S/R component.
volume_profile_source off Optional volume-profile input. Use auto, ticks, or m1_bars to enable it.
volume_profile_max_tick_window_days 1 Cap the raw-tick window, matching volume_profile_levels; longer auto windows use the disclosed M1 approximation.
volume_profile_max_ticks 50000 Cap the number of ticks pulled for volume profile.
volume_profile_max_m1_bars 20000 Cap M1 approximation work when volume profile is enabled.

The default min_source_families=2 requires independent agreement. Pass min_source_families=1 to include single-family clusters. Volume-profile computation is opt-in because it can require a much larger M1 history request than the other level families.

Output: levels sorted by score, each with a price, contributing source_families, and distance_pct from current price. Compact output keeps source_families; the contributing sources list appears at --detail standard or full.

Use support_resistance_levels for the structural S/R component on its own.

volume_profile_levels

Computes the volume-profile Point of Control (POC), Value Area High (VAH), and Value Area Low (VAL) from bounded raw ticks or an M1-bar approximation.

# Window by calendar range
mtdata-cli volume_profile_levels EURUSD --start "1 day ago" --end "now" --source auto --price-source mid --bucket-points 10 --json

# Window by lookback on a timeframe
mtdata-cli volume_profile_levels EURUSD --timeframe H1 --lookback 168 --source auto --bucket-points 10 --json

Choose one window mode. A calendar start cannot be combined with timeframe or lookback; end may anchor either mode.

Parameter Default Description
start / end — Calendar window for the profile.
timeframe — When set without lookback, the window defaults to the latest 200 completed bars.
lookback — Completed timeframe bars to include. The window follows actual market bars, so closures and weekend gaps do not consume the lookback.
source auto auto uses bounded ticks for short windows and M1 bars for larger ones; force with ticks or m1_bars.
price_source mid Price used per tick: mid, last, bid, or ask. M1 input cannot preserve quote sides and reports lch_equal_weight_proxy (volume split equally across each bar's low, close, and high; open omitted).
volume_source auto auto, candle real_volume/tick_volume, tick-snapshot volume_real/volume, or tick_count. tick_count is broker/provider tick rows and is rejected with --source m1_bars (use tick_volume or --source ticks). Snapshot volume is counted only on MT5 trade-change flags.
bucket_size / bucket_points / bucket_count — Choose exactly one price-bucket control (absolute size, points, or a target bucket count).
max_buckets 120 Upper bound on buckets.
value_area_pct 70.0 Percentage of volume that defines the value area (70 = 70%, the standard setting).
reference_price — Anchor for distance calculations (defaults to current price).
max_tick_window_days 1 Cap the tick window pulled. Natural relative windows within one second of the boundary count as in-budget; larger auto windows use M1 bars.
max_ticks 50000 Cap the number of ticks pulled.
max_m1_bars 20000 Cap the M1 bars pulled in approximation mode.

Output: poc, vah, val, and a value_area summary. Timeframe-derived queries also include bar_window with the resolved first open and last close. profile_source, source_decision, and volume_profile_accuracy identify whether auto used ticks or the M1 approximation and why. The top-level source remains the standard structured MT5 broker-provenance object. For latest timeframe-derived profiles, data_stale compares the completed window endpoint with one requested bar period (and exposes that threshold as stale_after_seconds); non-timeframe latest windows retain the five-minute observation-age threshold. Explicit historical start/end queries set data_stale to null (freshness_applicability=historical_query); judge those windows by coverage, gaps, and truncation instead of wall-clock age. --detail full adds the full levels histogram. To request a week of ticks, explicitly set --max-tick-window-days 7 and an appropriate --max-ticks cap.

Typical Workflow

  1. pivot_compute_points — mark today's formula pivots and R/S targets.
  2. support_resistance_levels — find the levels price has actually respected.
  3. volume_profile_levels — frame fair value (POC) and acceptance (value area).
  4. confluence_levels — surface zones where the methods above overlap, then trade plans around the highest-scoring, multi-family zones.

Caveats

  • Levels are descriptive analytics, not predictions or signals. Always confirm with price action and risk controls.
  • Volume profile on FX usually falls back to tick count; treat real_volume as available only when your broker provides it.
  • auto timeframe and auto source trade some precision for robustness; pin them explicitly for reproducible research.

See Also