About

About the maker.

Portrait of Jason Binet-Lafontaine, the developer behind AlphasWorth

I'm Jason Binet-Lafontaine, an individual doing business as AlphasWorth, based in Quebec, Canada. AlphasWorth is a one-person project.

I built these tools for my own trading. My frustration was a strategy builder and a live terminal that can quietly disagree: the same idea coded one way for the backtest and another way for the chart. So I wrote the tools I wanted for StrategyQuant X and MetaTrader 5, aiming for matching logic in the backtest and on the live terminal. Each product page says which platforms it covers.

That one frustration grew into a short list of rules I hold every module to, set out below. They matter more than any single feature: a backtest that quietly reads the future, or a table that hides how many things were tried, can look convincing and still mean nothing.

The tools started as part of my own workflow. I share them because other traders may want to save the same time, and because the sales help fund further work. AlphasWorth is an independent project. It is not endorsed by StrategyQuant or MetaQuotes. It is a StrategyQuant referral partner; see the disclosure.

Each product page lists its requirements and limits. Questions go to support. Phone and text numbers are on the contact page.

Trading leveraged instruments carries a high risk of loss. Software output is not investment advice and says nothing about future results.

Why these tools exist

Four rules, applied to every module

Most of the work in these modules is not the indicator or the model. It is making sure the number means what it appears to mean. This is how I try to do that.

Only what was knowable at the time

Alternative Data stamps every observation with the date it became public and joins on that date. News uses release times and never the released figures. The indicators step closed bars only, and a swing is confirmed only once bars on both sides of it have closed. Edge Finder reads a signal at one bar and enters at the next open.

The same logic in both programs, or a check that says otherwise

Each StrategyQuant X block has a MetaTrader 5 indicator written to the same specification, and the two are compared rather than assumed equal. NN Lab writes one training run to both platforms' model files. The Alternative Data feed service recomputes recent history and refuses to write if it disagrees with the research. How the checks work.

Count what was tried

The best of many tests is biased upward. Edge Finder adjusts every table for the number of cells tested (Benjamini and Hochberg q-values, and a permutation test across every time window searched), shrinks the headline figure for the number of candidates, and keeps your reserved bars out of the measurement.

Say what the tool cannot tell you

Order flow here is a proxy built from bars, not a depth-of-market feed. Wyckoff events are an interpretation. Edge Finder measures a tendency, not a profit. News is timing only. NN Lab selects a model and does not judge a strategy. Each product page has a limits section, and I would rather you read it before you buy than after.

Capabilities

What is in it for institutions and prop desks

Institutional Indicators is named for the vocabulary it implements (Wyckoff, order flow, TWAP, market structure), not for any institutional data. Alternative Data uses public series, not proprietary feeds. I make no claim that any fund, desk or prop firm uses these tools, and nothing here is a promise about results or about passing any firm's evaluation. Below is what each part does and the question it answers.

Wyckoff accumulation and distribution

A state machine over closed bars that reports climaxes, springs and upthrusts, signs of strength and weakness, phases A to E and the Creek and Ice range levels, with the stops and targets the range implies. The rules follow published Wyckoff texts; the events are one consistent interpretation, and no edge is claimed.

Institutional Indicators · manual

Order flow and TWAP

Cumulative delta with absorption, initiative and buy and sell pressure regimes, computed from bar volume and where price closed within the bar. It is a proxy, not a depth-of-market feed. TWAP is a rolling time-weighted average price usable as a level, with rising, falling and cross blocks.

Institutional Indicators · other families

Market structure and levels

Breaks of structure, changes of character, liquidity sweeps, fair-value gaps and order-block retests on confirmed swings, plus support and resistance, round numbers, session levels and the opening range. It deliberately does not label Wyckoff events, so it is not a second Wyckoff model.

Institutional Indicators

Patterns and moving-average maths

Fourteen multi-candle formations as strict true/false blocks, bar and chart patterns, and a configurable moving average (ALMA, DEMA, ZLEMA, T3, Kalman, SMA, EMA, SMMA, LWMA) with slope and context blocks.

Institutional Indicators

Alternative data

Eighteen features built from non-price series, for 24 instruments: the VIX and an instrument's own implied-volatility index where one exists, the broad dollar index, CFTC positioning, EIA petroleum inventories and US macro series, as builder blocks. They come from free public sources (FRED, CFTC and EIA, with your own free keys), each joined on the date it became public. Whether any series carries edge on your market is the builder's question to answer.

Alternative Data · manual

News blackout and timing

Blocks built from the economic calendar inside your MetaTrader 5 terminal: out five minutes before a release and nothing new until fifteen after, or minutes to the next release as a quantity. The calendar's timezone is measured on every import, only schedule is used and never outcome, and with no calendar the all-clear is false. It is a timing tool and does not promise to keep an account inside a loss limit.

News · manual

Neural networks (NN Lab)

Trains MLP, convolutional, LSTM and CONVLSTM networks inside StrategyQuant X on built-in features, SQX blocks or Alternative Data columns, and writes one model for the builder and one for MetaTrader 5 holding the same weights. The tail you reserve for out-of-sample testing is dropped, not scored, and a purge gap separates it from training.

NN Lab · manual

Edge Finder

Measures which side leans, which hours and weekdays lean, trend against reversion, and which blocks carry a tendency, before you run a build, then writes the findings into a Build task and reads the task back to confirm. It is a bias detector, not a backtest or a profit forecast, and the same data and settings give the same table.

Edge Finder · manual

GapSmith: data hygiene

Tick-precision simulation builds its price path from M1 bars, so a missing minute has no path at all. GapSmith fills genuine gaps (never weekends, holidays or closures) into a separate _patched symbol with an audit log, flat by default, with an optional synthetic fill that is marked as such. It does not recover prices that were never recorded.

GapSmith · manual

SpecSmith: specification and timezone

Compares the tick size, point value, swap, commission and sessions StrategyQuant X holds with what your MetaTrader 5 terminal reports, applies the rows you tick, and can undo them. Its timezone check tests the zone you typed against the weekly forex reopening at 17:00 New York time, and only reports.

SpecSmith · manual

NameSmith

Renames the strategies you select from a template such as Breakout_SP500_H4_001, with a separate counter per ticker and timeframe, a preview and an undo, so a databank can be audited. It renames; it does not rank or judge anything.

NameSmith · manual

The MetaTrader 5 mirror

A compiled MetaTrader 5 indicator for each indicator family, plus feed services for alternative data and news, so an exported strategy reads the same logic and the same files the backtest used. MetaTrader 5 is the only supported export platform for these modules.

The workflow · Requirements

Trading leveraged instruments carries a high risk of loss. Software output is not investment advice and says nothing about future results.