FreeOpen sourcePine Script v6
Fix round-number levels before observing reactions. Track whether later closes return to the approach side or pass through, then compare with two shifted grids under matched prior conditions.
ROUND levels are integer multiples of a zero-origin interval, not selected historical reaction points. References A/B share the spacing and are matched by time, approach side and prior range/trend context. Negative differences, disagreement and insufficient samples are legitimate findings.
Default colors are cyan for ordinary levels, yellow every fifth interval, mint for the enclosing lower level and pink for the upper. Enclosing roles override major color. Badges retain exact level prices; R/T/U mark known ROUND outcomes. Optional dotted A and dashed B lines can stay hidden while research continues.
For interval I, levels are 0, I, 2I and so on. Default 200 ticks equals 0.200 with a 0.001 tick or 0.00200 with a 0.00001 tick. A shifts by a tick-rounded 37% of I; B takes the complementary ticks. Automatic preview depends on initial history, so use Manual price to lock its chosen interval.
A candle range must intersect a level band while the prior close has adequate clearance on one approach side. Default band half-width is 4% of interval, minimum one tick; clearance defaults to 12%, at least one tick beyond the band.
If an opening gap crosses an entire band in any grid, or one candle touches multiple level bands in any grid, all three reject new contacts on that candle. Active observations continue. Overlapping configured contact bands also pause comparison.
Default response distance is 20% of interval, optionally a prior finite mean-range multiple, at least one tick beyond the band. It freezes at contact. Exclude the contact candle: subsequent closes at or beyond the approach-side threshold count toward R, opposite-side closes toward T, with two consecutive closes by default.
The first confirmed R or T is fixed; each counter resets on its own failed condition. Last-horizon-bar confirmation precedes U. Even early results wait the full default 12-bar horizon before comparison. Other unresolved contacts do not impose a busy state or outcome-dependent cooldown.
Epoch-anchored clock blocks default to 128 chart periods. Within each block, side and context stratum, match first ROUND/A/B contacts, then second contacts, up to eight each by default. Reject triplets exceeding half-block time spread without rematching to preferred partners.
Prior finite mean true range divided by interval is low below 0.25, middle from 0.25 to below 0.75, high at or above 0.75. Prior close minus the close ten bars earlier defines trend with an inclusive ±0.5 prior-range neutral band. Defaults give two sides × three range × three trend strata.
Use complete blocks whose contacts have all received the full horizon, default latest 24 mature blocks. Exclude blocks cut by history or study dates. Start is inclusive, end exclusive; admitted contacts still mature after the end date. Complete empty blocks do not become fabricated trades.
Use a standard time-based chart and check the Interval price value first. The default is 200 minimum ticks, not 200 pips. Start with Compact and specify interval, dates and observation rules before interpreting differences.
Open the public Round Number Levels – Response and Control page and add it to TradingView. No SG Group account is required.
Overlapping lines, bands, backgrounds or tables can be duplicate drawings of the same kind of information. Never treat visual overlap by itself as stronger evidence or signal agreement. Base levels, major emphasis, enclosing roles, half guides and A/B references serve different functions. Other horizontal tools can obscure prices or make related drawings look like independent evidence.
Compact has 14 rows, Standard 22 and Research 32. Read interval and denominator before rates, reference differences, disagreement and window coverage.
Interval is spacing in price units. Lower/Upper are exact enclosing levels; distances use ticks and Cell position locates price within the interval. Displacing badge boxes does not move their printed value or level line. Digits-only badges retain roles through colors and tooltips.
N each grid shows equal matched counts for ROUND, A and B, not storage limits or visible markers. Return/Through/Unresolved compare ROUND with the arithmetic mean of A/B shares. U remains in the denominator; tooltips show individual reference counts and shares.
R delta A / B subtracts each reference return share from ROUND. A 40% versus 35% share gives +5 percentage points, not a 5% relative gain or future probability. Mean averages the two differences; SPLIT exposes opposite signs even when the mean is positive.
Default main deltas require 30 triplets, five per side and three contributing blocks. Counts and outcome shares can remain visible when deltas are withheld. Early/Late uses contiguous halves of the retained window with smaller denominators and separate guards, not a held-out test.
Standard adds all mature contacts, matched coverage, over-cap counts, contributing blocks, spread and window coverage. Research adds exact enclosing-level matched history, side counts, separate A/B outcomes, shared exclusions and reproduction details. Matching can retain only part of the contact population.
Markers show eligible ROUND outcomes before matching, while the panel uses matched fully mature triplets. When several resolve on one candle, the oldest contact represents that candle and + indicates more. Inspect the R/T/U counts in its tooltip. Visible-marker counts cannot reconstruct the denominator.
Live enclosing levels and shading may change intrabar; observations and results use confirmed bars. The matched panel waits for closed mature blocks and need not change every candle. Showing only the latest unresolved thresholds does not stop other observations.
RN6_ outputs retain protocol 6001, settings, clock window, actual N, differences, mean, sign status and a non-cryptographic OHLC checksum. Each grid exports contact time, tick level, side, fixed threshold, result, first-confirmation age, stratum and ordinal at contact plus effective H. Unmatched/over-cap records remain; negative ordinal identifies a partial/date-cut block exclusion.
Base round grid / shifted A / complementary B, equally spaced but distinct levels.
Return to approach side / through to opposite side / unresolved at horizon. R/T require consecutive closes.
Additional outcomes share the bar; the oldest contact is representative. Tooltip provides counts.
Equal matched count per grid, not visible-marker count or capacity.
ROUND-minus-A/B return-share differences and their mean, in percentage points.
Both differences positive / both negative / opposite signs / one or both exactly zero; not significance labels.
Contiguous early/late halves with their own smaller denominators and guards.
Automatically previewed spacing, history-sensitive until manually locked.
Low sample / partial coverage / fixed spacing and displayed-window coverage; not an optimization instruction or performance certificate.
Numeric export prefix / research protocol identifier.
The labels below match the published script’s English Inputs panel. Defaults reflect the public source at the time of review and may change in a later TradingView update.
On small screens, each input is reformatted as a readable stacked card.
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Spacing mode | Manual ticks / options: Manual ticks, Manual price, Auto preview (prior range), Auto preview (price %) | Manual ticks fixes a tick interval; Manual price fixes a price interval. Automatic previews choose a 1–2–5 spacing once from initial usable prior range/price context, so lock the displayed value manually for repeatability. |
Fixed interval (minimum ticks) | 200 / min 1, max 1000000000 | Fixed spacing in minimum ticks, default 200, not 200 pips. A 0.001 tick yields 0.200; a 0.00001 tick yields 0.00200. |
Fixed interval (price units) | 0.2 / min 1e-9, max 1000000000000 | Fixed price-unit interval, default 0.2, rounded to whole minimum ticks. The zero-origin grid is not fitted to earlier reactions. |
Prior mean true-range length | 50 / min 2, max 500 | Finite simple mean true-range length, default 50, not recursively seeded Wilder ATR. Research uses the previous bar’s value. |
Preview interval / prior mean range | 3 / min 0.25, max 30 | Prior mean-range multiple for automatic spacing preview, default three; unused in manual modes. |
Preview interval / prior price (%) | 0.25 / min 0.001, max 20 | Prior-price percentage used by price-based spacing preview, default 0.25%. |
Major level every N intervals | 5 / min 2, max 20 | Major-line emphasis every N intervals, default five. Drawing only, not a cohort-selection rule. |
Study start (inclusive) | 2025-01-01 00:00 UTC | Inclusive study start, default 2025-01-01 00:00 UTC. Earlier finite-range/trend context is still needed. |
Study end (exclusive) | 2100-01-01 00:00 UTC | Exclusive end, default 2100-01-01 00:00 UTC. Existing contacts still receive their full horizon; date-cut blocks do not enter the matched sample. |
Require full start-date context | Off | Requires loaded history across the requested start plus sufficient preceding context. Otherwise research stays off. When disabled, the initial partial block is excluded and coverage is disclosed; grid drawings remain available. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Enable three-grid research | On | Enables three-grid research for ROUND and references A/B, independently of the price-grid display. |
Contact half-width (% of interval) | 4 / min 0.1, max 10 | Contact-band half-width, default 4% of interval with a one-tick minimum. Contact tests intersection with the candle’s high-low range. |
Prior-close clearance (% of interval) | 12 / min 1, max 45 | Prior-close clearance from the level, default 12% of interval, at least one tick beyond the contact band. |
Response distance | Interval fraction / options: Interval fraction, Prior mean range | Response-distance basis: interval fraction or prior mean range. The chosen distance is frozen at contact. |
Response threshold (% of interval) | 20 / min 1, max 200 | Interval-based response distance, default 20%, with a floor at least one tick beyond the contact band. |
Response threshold / prior mean range | 0.6 / min 0.05, max 10 | Prior mean-range multiple for the response threshold, default 0.6; subsequent volatility does not move it. |
Consecutive closes to confirm | 2 / min 1, max 10 | Consecutive qualifying closes, default two. Each side’s counter resets whenever its own condition fails. |
Full observation horizon (bars) | 12 / min 1, max 100 | Full horizon after contact, default 12, at least the confirmation count. Early outcomes wait until full maturity for comparison; concurrent contacts are not blocked. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Reference A / interval | 0.37 / min 0.15, max 0.85 | A offset as interval fraction, default 0.37 rounded to ticks. B uses the exact tick complement, normally about 0.63. Overlapping contact bands pause comparison. |
Clock block (chart periods) | 128 / min 16, max 2048 | Epoch-anchored elapsed-time block length, default 128 chart periods; on 5-minute charts this is 640 minutes, not 128 actually traded bars. |
Maximum matched time spread (% of block) | 50 / min 1, max 100 | Maximum earliest-to-latest contact-time spread in a triplet, default half a block. Failed triplets are not rematched. |
Match on prior context | Side + range + trend / options: Side only, Side + range, Side + range + trend | Matches on approach side, optionally prior range and trend bins. Within each block/stratum, pair first contacts, then second contacts, and so on, without using outcomes. |
Range/interval: low boundary | 0.25 / min 0.01, max 10 | Low boundary for prior mean range divided by interval, default 0.25; values below it are low. |
Range/interval: high boundary | 0.75 / min 0.02, max 20 | High boundary, default 0.75; with defaults, 0.25–<0.75 is middle and ≥0.75 is high. |
Prior trend lookback (bars) | 10 / min 1, max 500 | Prior trend lookback, default 10: preceding close minus the close ten bars before that. |
Prior trend neutral band / mean range | 0.5 / min 0, max 20 | Neutral trend band, default ±0.5 prior mean range, including both boundary values. |
Retained mature clock blocks | 24 / min 4, max 64 | Retained mature clock blocks, default 24, shared by all grids. Partial blocks are excluded; complete empty blocks are allowed without fabricating observations. |
First contacts per grid / stratum / block | 8 / min 1, max 12 | Outcome-blind first-K matching cap per grid/stratum/block, default eight. Later contacts are still tracked/exported and cap exclusions counted. |
Minimum matched triplets for deltas | 30 / min 3, max 1000 | Minimum matched triplets for main deltas, default 30. Counts and outcome shares remain visible below the guard. |
Minimum triplets per approach side | 5 / min 1, max 200 | Minimum triplets per approach side, default five. |
Minimum contributing clock blocks | 3 / min 1, max 64 | Minimum contributing blocks, default three. A display guard, not evidence of significance or independence. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Base levels on each side | 6 / min 2, max 24 | Base drawing levels on each side, default six; does not select research observations. |
Line length behind current bar | 260 / min 20, max 4000 | Line segment length behind the current bar, default 260 bars. |
Line end to the right (bars) | 14 / min 2, max 100 | Right-side line endpoint, default 14 bars. |
Line extension | Segment / options: Segment, Right, Both | Segment, right extension or both directions. A visual projection of the fixed grid, not historical signals. |
Reduce distant-line clutter | On | Reduces distant-line clutter. |
Drawing radius / current mean range | 12 / min 2, max 100 | Drawing radius in current mean ranges, default 12. Enclosing base levels remain and research observations are not removed. |
Show level price badges | On | Shows level-price badges. |
Show half-interval guides | Off | Optional half-interval guides, hidden for odd tick intervals to avoid uneven subdivision. Never researched. |
Show reference A lines | Off | Shows dotted A lines; research continues when hidden. |
Show reference B lines | Off | Shows dashed complementary B lines; visibility does not control research. |
Shade the current price cell | On | Shades the cell enclosing current price. |
Follow the live price for visual context | On | Allows enclosing lines, distances and shading to follow live price intrabar. Observations, outcomes and research use confirmed bars. |
Show latest unresolved round-contact thresholds | On | Shows thresholds for the latest unresolved ROUND contact only. Other contacts continue; these are research thresholds, not stops or targets. |
Mark completed round-grid outcomes | On | Shows R/T/U on the confirmed outcome bar, never backdated to contact. |
Maximum outcome markers | 60 / min 0, max 250 | Maximum visible outcome markers, default 60; zero hides them without truncating the ledger. |
Base line width | 1 / min 1, max 4 | Base-line width, default one. |
Major line width | 2 / min 1, max 4 | Major-line width, default two. |
Enclosing line width | 3 / min 1, max 5 | Enclosing-line width, default three. |
Price badge placement | Left of segment / options: Left of segment, Right of segment | Places badges left or right of segments. Default left placement helps avoid the right-hand panel. |
Show response-threshold captions | Off | Shows response-threshold captions; off by default. |
Minimum marker gap at the same level | 12 / min 0, max 100 | Minimum spacing from the latest drawn marker at the same level, default 12 bars. Drawing-only thinning; concurrent outcomes and full-horizon records remain. |
Round-number font size (points) | 13 / min 10, max 36 | Base round-number badge font size, default 13 points, independent of panel and outcome markers. |
Major / enclosing font size (points) | 14 / min 10, max 40 | Major/enclosing font size, default 14 points, at least the base badge size. |
Round-number badge appearance | Solid level color / options: Solid level color, Dark badge | Solid level color uses opaque level-colored badges with automatic black/white contrast. Dark badge uses white on navy without recoloring lines or panel. |
Price digits only in round badges | On | Shows digits only by default; disable to restore L/U and other prefixes. Roles remain in colors, tooltips and panel. |
Monospace round-number digits | On | Uses monospace round-number digits. |
Round-badge horizontal gap (bars) | 12 / min 0, max 150 | Horizontal badge gap from the segment endpoint, default 12 bars. Leaders keep the true endpoint; right placement may collide with the panel. |
Price badge offset (ticks) | 2 / min 0, max 100 | Badge-box vertical displacement, default two ticks. Printed price and line remain exact; enclosing badges move outward. |
Alternate badge offset direction | On | Alternates displacement for non-enclosing badges. Off uses upward displacement for other levels; enclosing lower/upper badges remain outward. |
Outcome marker placement | Above/Below bar / options: Above/Below bar, On level | Above/Below bar clears the outcome candle and selected preceding bars; On level uses the level. Time stays on the outcome bar and no future bars are used. |
Outcome marker offset / mean range | 0.3 / min 0.05, max 3 | Additional clearance in mean-range units, default 0.3, fixed at outcome confirmation. It does not alter research thresholds. |
Marker clearance lookback (bars) | 3 / min 1, max 25 | Clearance lookback includes outcome and preceding bars, default three total. Future candles may later enter that area. |
Outcome marker size | Normal / options: Small, Normal, Large | Outcome-marker size: Small, Normal or Large. |
Connect displaced badges to exact levels | On | Connects displaced badges to their exact levels. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Show research panel | On | Shows the research panel. |
Panel detail | Compact / options: Compact, Standard, Research | Compact has 14 rows, Standard 22 and Research 32. Detail exposes coverage and separate references without changing observation or matching. |
Panel location | Top right / options: Top right, Middle right, Bottom right, Top left, Bottom left | Panel position: top/middle/bottom right, top left or bottom left. |
Panel label size (points) | 14 / min 10, max 22 | Panel label size, default 14 points. |
Panel value size (points) | 15 / min 10, max 24 | Panel value size, default 15 points. |
Show panel icons | On | Shows supplementary panel icons. |
Right-side panel inset (%) | 2 / min 0, max 15 | Right-side panel inset, default 2%. |
Print reproducibility recipe to Pine Logs | Off | Writes a reproduction recipe at the last confirmed historical bar in personal editor copies only. Published scripts cannot create Pine Logs; public users can export protocol fields. No upload or network call. |
| Input label | Default / options | What it controls and when to adjust it |
|---|---|---|
Base levels | rgba(35,215,255,1) | Base-level color, cyan by default. |
Major levels | rgba(255,202,65,1) | Major-level color, yellow by default. |
Enclosing lower level | rgba(78,245,183,1) | Enclosing lower level, mint by default; overrides major-level color. |
Enclosing upper level | rgba(255,100,173,1) | Enclosing upper level, pink by default; overrides major-level color. |
Shifted comparison | rgba(180,139,255,1) | Shifted-reference color, purple by default. |
Return outcome | rgba(95,240,196,1) | Return-result color, mint by default; not a buy signal. |
Through outcome | rgba(255,146,90,1) | Through-result color, orange by default; not a sell signal. |
Neutral / unresolved | rgba(163,179,199,1) | Neutral/unresolved color, blue-grey by default. |
Panel and badge background | rgba(15,23,38,1) | Panel and badge background color. |
Panel text | rgba(237,244,255,1) | Panel text color. |
Cell shade transparency | 95 / min 85, max 100 | Cell-shading transparency, default 95; larger values are fainter and 100 is transparent. |
Create the required alerts in TradingView. They describe observations, not trade instructions. Recreate existing alerts after changing research settings so they use the intended configuration.
This tool is for analysis and visualization. It is not investment advice, a trade recommendation, a forecast, or a promise of profit or accuracy. Results depend on inputs, symbol, timeframe, and data feed. You remain responsible for every trading decision.
Shifted references may themselves be round at finer scales. They are not randomized controls or exact counterfactuals. Broad strata leave different paths and unmeasured conditions; Both + proves neither a causal effect, future reversal nor profitability.
Horizons can overlap. First-K caps, context boundaries, interval and period affect the studied subset. Changing settings after seeing results creates selection risk; small, negative or inconsistent differences remain valid findings.
Automatic interval selection depends on initial usable history. Manual spacing and identical sufficient prior OHLC support finite-context and epoch-block reproducibility, but feed revisions, adjustments, sessions and missing bars can change results.
Full-start context requires history crossing the start with warm-up. Otherwise the initial partial block is excluded. Fixed / full describes the displayed window, not the completeness of all requested history or statistical validity.
Default glyph clearance uses the outcome candle plus two previous candles and a mean-range margin. Later candles may enter that area. Font, spacing, retention and radius controls affect presentation, not research thresholds or membership.
Non-standard and tick charts retain the grid without comparison research. Blocks are elapsed time, not sessions or traded-bar counts. Weekly/monthly periods use platform seconds conversion rather than calendar-aligned boundaries.
Emission at effective H bars after contact allows at most one complete record per grid/bar, even if early outcomes coincide. Preserve source, settings and native OHLC including prior context and subsequent horizon. The checksum is not a cryptographic signature. Pine Logs recipes work only in personal editor copies, not the published script.
This indicator observes fixed-level contacts and confirmed-close responses. It models no entry, stop, target, spread, fee or position size and places no orders. R is not automatically profit and T is not automatically loss.
Clarify units, denominators and information timing.
No. Multiply 200 by the symbol minimum tick. With 0.00001 ticks the interval is 0.00200. Check the displayed price value.
No. Line visibility is separate from Enable three-grid research.
No. It marks the later bar when consecutive closes confirm return, without backdating to contact.
Markers are thinned pre-matching ROUND outcomes; the panel uses fully mature matched triplets. Scope and timing differ.
It states that both observed return-share differences are positive, not that they are causal, significant, predictive or profitable.
It is a free open-source TradingView indicator with no external service or data fetch requirement. Research requires standard time-based charts.
This guide is grounded in SG Group’s live catalogue, the script’s published TradingView description, and the user-facing inputs in its open source. Labels and defaults can change after a script update.
