NomadTradingFX · The Nomad Trader

Parameters Reference

Every setting in the Inputs tab, what it does, and when to touch it. Companion to the methodology manual.

00  Orientation

Start here

The defaults are chosen to work out of the box on liquid instruments. Most users never change anything. The settings are organised into nine labelled groups in the Inputs tab; you can expand the ones you care about and leave the rest collapsed.

If you change only three things 1. Min tradeable leg size (Swing Detection), set near the size of moves your strategy actually trades. 2. The six factor weights (Factor Weights), if you want to shape what "good" means or fit it to your own robot. 3. Good / fair score (Grading), to make the verdict stricter or looser. Everything else is fine on defaults.

A note on units used below: ATR means the setting is measured in multiples of Average True Range, so it adapts to each instrument and timeframe automatically. Points are raw price points (ten points to a pip on a 5-digit feed). Bars are candles on the timeframe being measured.

01  Group

Analysis Windows

How much recent price each measurement reads. Bigger windows are steadier and slower; smaller windows are more reactive and noisier.

Condition windowInpEvalBars
default 200range 120–350integer

The number of recent bars examined to judge current conditions, where swings are detected and most factors are read. Think of it as how far back "now" reaches.

▲ higher: steadier reading, needs a longer move to register, slower to react
▼ lower: reacts faster to fresh shifts, but noisier and easier to fool
Market Meanness lookbackInpMMIPeriod
default 150range 80–250integer

How many bars feed the trend-versus-random factor. A longer lookback gives a more statistically stable read on whether price is genuinely going somewhere.

▲ higher: more stable trend/random read, slower to turn
▼ lower: more responsive, noisier
Fractal Dimension lookbackInpFDIPeriod
default 30range 20–60even integer

How many bars feed the smooth-versus-jagged factor. This one is short by nature because it captures local texture. Odd values are rounded down to keep it even.

▲ higher: smoother, more averaged texture read
▼ lower: more local and reactive
InteractionThe amount of downloaded history needed scales with the largest of these three windows. Pushing them very high makes the higher timeframes slower to leave LOADING on thin-history symbols.

02  Group

Self-Calibration Baseline

This group controls the heart of the tool: how each factor is normalised against that timeframe's own history. A score near 50 means "normal for this market"; high means "unusually good for this timeframe right now."

Baseline historyInpBaselineBars
default 1200range 600–3000integer

How far back the engine looks to learn what is "normal" for each factor on each timeframe. This is the reference the current reading is compared against.

▲ higher: steadier, more representative norm, needs more history downloaded
▼ lower: adapts faster to a changed regime, noisier norm
When to changeIf the higher timeframes (especially D1) sit on LOADING because your broker holds little history, lower this to 600–800 so the baseline can form from what is available.
Baseline samplesInpBaselineSamples
default 16range 12–40integer

How many windows across that history are sampled to estimate each factor's normal range. More samples give steadier statistics, at a little more processing.

▲ higher: smoother, more reliable baseline, slightly heavier
▼ lower: lighter, but shakier z-scores
When to changeRaise it toward 24–32 if scores feel jumpy or unstable from one moment to the next.
Baseline refresh secondsInpBaselineSec
default 120range 60–600seconds

How often the baseline is rebuilt. The baseline moves slowly, so it is cached and refreshed occasionally rather than every tick.

▲ higher: lighter on CPU, baseline lags a regime change a little more
▼ lower: baseline tracks change sooner, more CPU
z-score steepnessInpNormSteep
default 1.2range 0.8–2.0decimal

The contrast dial. It sets how sharply a deviation from normal translates into a high or low factor value. This is the most useful single knob for the overall feel of the scores.

▲ higher: harsher, more separation, GOOD and poor appear more readily
▼ lower: gentler, scores hug the middle, fewer extremes
When to changeIf readings feel flat and everything sits around fair, raise it toward 1.5. If the panel swings too violently, lower it toward 1.0.

03  Group

Swing Detection

How price is broken into tradeable legs. These two settings define what counts as a swing and what counts as worth trading, both measured in ATR so they port across instruments.

Swing reversal thresholdInpReversalATR
default 0.75range 0.4–1.5x ATR

How far price must retrace, in ATR, before a swing is considered to have reversed. This decides how finely the engine carves price into legs.

▲ higher: only larger reversals count, fewer and bigger legs
▼ lower: detects smaller swings, more and finer legs
Min tradeable leg sizeInpLegMinATR
default 1.5range 1.0–3.0x ATR

The smallest leg, in ATR, that counts as a real trading opportunity. Legs below this are ignored. This is the single most important setting to align with your own system.

▲ higher: only big moves count, stricter, favours large clean trends
▼ lower: counts smaller moves as opportunities, lets faster timeframes qualify
Match it to your strategySet this close to the size of the moves your robot or your discretionary entries actually capture. Keep it at or above the reversal threshold. It also feeds the cost gate, since it defines the average leg compared against the spread.

04  Group

Factor Weights

The relative importance of the six properties the engine measures. They are auto-normalised, so only their ratios matter; they do not need to sum to anything. Raise one to make that property count for more, or set it to zero to switch a factor off. The defaults are a balanced, generic trend profile.

SettingDefaultQuestion it answers
InpW_MMI0.22Trend versus random. Raise for strict trend-only conditions.
InpW_FDI0.18Smooth versus jagged. Raise if your system dislikes choppy texture.
InpW_SNR0.15Swing versus noise. Raise if entries and exits suffer when noise is high.
InpW_Clean0.13Trend versus whipsaw. Raise if your system bleeds in back-and-forth chop.
InpW_Opp0.14Setup frequency. Raise if you want more trades or faster timeframes to compete.
InpW_Prog0.18One-way progress versus round-trip. Raise if you need follow-through, not V-shapes.
The real useOut of the box these measure generic trend quality. To make GOOD mean "my system profits here," turn on the calibration log, gather the factor readings beside how your system actually performed per timeframe, and fit these six weights to that outcome. The methodology manual walks through the procedure.

05  Group

Cost Gate

A safety floor that demotes a timeframe to poor when its moves are too small to clear trading costs. It is the only absolute check in the tool, and it never taxes the score otherwise, so faster timeframes are not punished simply for being faster.

ATR periodInpATRPeriod
default 14range 10–20integer

The Average True Range period used for all ATR scaling throughout the tool, including the swing thresholds. The standard value suits almost everyone.

▲ higher: steadier volatility yardstick
▼ lower: adapts to volatility shifts sooner
Round-turn commissionInpCommissionPts
default 0.0pointsdecimal

Your broker's round-turn commission expressed in points, added to the spread so the cost gate reflects true cost. Leave at zero on commission-free spread accounts.

Raw and ECN accountsSet this to your real commission in points so the gate judges cost honestly. Otherwise the lower timeframes will look cheaper to trade than they are.
Minimum leg-to-cost ratioInpMinLegCost
default 2.5range 1.5–5.0decimal

How many times larger than the round-turn cost the average tradeable leg must be. Fall below this and the timeframe is vetoed to poor, regardless of how good its conditions look.

▲ higher: stricter, demands moves comfortably clear the spread
▼ lower: lets marginal-cost timeframes through

06  Group

Grading

Where the 0 to 100 score is cut into GOOD, fair and poor. Since scores centre near 50 for a normal market, the defaults treat clearly-above-normal as good and clearly-below as poor.

Score for GOODInpGoodScore
default 60range 55–75decimal

The score at which a timeframe earns a GOOD rating and becomes a deployment candidate.

▲ higher: GOOD becomes rarer and more selective
▼ lower: GOOD appears more easily
Score for fairInpFairScore
default 40range 30–50decimal

The line between fair and poor, and the bar a timeframe must clear before the hero will recommend it instead of showing STAND ASIDE. Keep it below the GOOD score.

▲ higher: more timeframes fall to poor, stricter stand-aside
▼ lower: more timeframes count as fair

07  Group

Behaviour

How responsive and how steady the panel is.

HysteresisInpConfirmBars
default 1range 1–5recalcs

How many consecutive recalculations a new rating must hold before it is committed. This stops ratings flickering when a score sits right on a band boundary. The count is in recalculation cycles, not chart bars.

▲ higher: steadier ratings, slower to switch
▼ lower: more responsive, may flicker at a boundary
Recalculation secondsInpRefreshSec
default 2range 1–10seconds

How often the whole panel recalculates. It runs on a timer, which is why it keeps updating even when the market is closed and no ticks are arriving.

▲ higher: lighter on CPU, less responsive
▼ lower: snappier updates, more CPU

08  Group

Calibration Log

A diagnostic feed for when you want to see inside the score, mainly while fitting the weights to your own system. Off by default.

Print factor breakdownInpDebugLog
default falseon / off

When on, each timeframe's six normalised factors, its cost ratio, and its final score are printed to the Experts tab. This is your window into why a timeframe scored the way it did.

Turn it on to calibrateRun with this enabled across the markets and periods where you know how your system performed, then fit the weights to separate the timeframes it did well on from the rest. Turn it off for normal use.
Seconds between printsInpDebugSec
default 10range 5–60seconds

Throttles how often the breakdown is written, so the log stays readable rather than scrolling past.

09  Group

Dashboard

Purely cosmetic. Size and position of the on-chart panel.

Panel sizeInpScale
default 2.0range 1.0–3.0decimal

A single multiplier for the whole panel, fonts and layout together. Raise it on large or high-resolution screens, lower it to tuck the panel into a corner.

▲ higher: larger panel
▼ lower: smaller panel
Panel positionInpPanelX · InpPanelY
default 16 · 24pixels

Distance of the panel from the top-left corner of the chart, across and down. Move it to keep it clear of your price action.

10  How to

Tuning recipes

Make GOOD rarer and more selective

Raise Score for GOOD to 65–70, or raise z-score steepness toward 1.5.

Everything reads middling and flat

Raise z-score steepness toward 1.5. Confirm the higher timeframes have finished loading their history.

See smaller swings on a quiet pair

Lower Swing reversal threshold and Min tradeable leg size together.

Match it to my own system

Set Min tradeable leg size to your typical traded move, set the commission, then fit the six weights with the calibration log on.

Higher timeframes stuck on LOADING

Lower Baseline history to 600–800, and scroll the D1 or H4 chart back once to pull its history.

Ratings flicker between bands

Raise Hysteresis to 2 or 3.

Panel is too big or in the way

Lower Panel size, and adjust the panel position.

Reduce processing load

Raise Recalculation seconds and Baseline refresh seconds, and lower Baseline samples.

11  Reference

Quick reference

SettingDefaultGroup
Analysis Windows
InpEvalBars200Condition window in bars
InpMMIPeriod150Market Meanness lookback
InpFDIPeriod30Fractal Dimension lookback (even)
Self-Calibration Baseline
InpBaselineBars1200History normalised against
InpBaselineSamples16Sample windows for the baseline
InpBaselineSec120Seconds between refreshes
InpNormSteep1.2z-score steepness (contrast)
Swing Detection
InpReversalATR0.75Swing reversal threshold, x ATR
InpLegMinATR1.5Min tradeable leg, x ATR
Factor Weights
InpW_MMI0.22Trend vs random
InpW_FDI0.18Smooth vs jagged
InpW_SNR0.15Swing vs noise
InpW_Clean0.13Trend vs whipsaw
InpW_Opp0.14Setup frequency
InpW_Prog0.18One-way vs round-trip
Cost Gate
InpATRPeriod14ATR period for scaling
InpCommissionPts0.0Round-turn commission, points
InpMinLegCost2.5Min leg-to-cost ratio or veto
Grading
InpGoodScore60Score for GOOD
InpFairScore40Score for fair
Behaviour
InpConfirmBars1Hysteresis in recalcs
InpRefreshSec2Seconds between recalcs
Calibration Log
InpDebugLogfalsePrint factor breakdown
InpDebugSec10Seconds between prints
Dashboard
InpScale2.0Panel size multiplier
InpPanelX16Panel X from left
InpPanelY24Panel Y from top