Glassnode logo

Glassnode Alpha Lab Beta

Find Your Edge.

Turn metrics into actionable signals. Alpha Lab searches Glassnode's data for simple, readable trading rules, then shows you the evidence behind each one.

  • 1,000+ metrics
  • Walk-forward validation
  • Human-readable rules

What It Is

Turn Glassnode Metrics Into Testable Trading Signals in Minutes.

The Problem

On-chain, spot, futures, options, order books. No one else measures the market from this many angles.

But breadth creates its own challenge. Among thousands of metrics, only a handful may matter for a specific investment decision. Finding, testing, and turning them into repeatable rules takes time, code, and quantitative expertise.

The advantage is already in the data. The hard part is finding it and turning it into something actionable.

The Answer

Alpha Lab starts with your question, not the data. Say what you want to identify, pick the metrics to explore, and it searches for the combinations that have historically been most useful.

You get heuristics, not black-box predictions. Rules of one or two conditions, in plain language.

How To Use It

Three Modes, One Engine.

Three ways into the same research engine. Different levels of control, one evaluation framework.

Go from question to answer

Easy mode

From a question to ranked heuristics. Easy Mode abstracts away the optimization and machine-learning workflow.

  • Choose an asset, a direction and the metrics to scan
  • Results arrive as ranked, readable signal cards
  • Refine the metric set and re-run to compare

Inspect every step

Advanced mode

Every decision behind the optimization and machine-learning workflow. More flexibility.

  • Pick the transformations: z-score, RSI, moving average, volatility, percentile
  • Paint your own target zones directly onto price history
  • Compare models, inspect feature influence, test threshold sensitivity

Put autoresearch to work

Agents mode

An MCP server that lets compatible AI agents drive the research loop.

  • Same catalogue and optimization engine as the visual interface
  • Give an agent an objective and let it run, evaluate and refine
  • Hand the strongest findings back to the UI for inspection

How It Holds Up

A Rule You Can Read, and Check.

Historical performance is easy to overstate, so skepticism is built into the workflow. Every rule arrives with the evidence needed to judge it.

An Example Result

Percent Entities in Profit RSI(30) > 51.8andNUPL Volatility(30) 1.71

Two conditions, in the metrics you already know. Nothing to decode, and nothing hidden behind a model score.

Out of sample · 1,090 days the search never trained on

Out-of-sample equity curve: the example rule against Bitcoin buy and holdOver the 1,090 days held out of training, from August 2023 to August 2026, the rule grows an initial 1.0 to 2.96 while buy and hold reaches 2.48, with a shallower drawdown along the way.1.01.62.84.8Aug 2023May 2024Feb 2025Nov 2025Aug 2026The ruleBTC buy & hold
Equity, log scale, both rebased to 1.0 at the start of the held-out window.
 The ruleBuy & hold
Sharpe ratio1.360.89
Return, annualized+44%+36%
Max drawdown−20%−53%
Time in market42%Always

A Research Process Built on Sound Methodology

Time-aware validation

Validation runs in stages. Walk-forward cross-validation trains models on earlier data and evaluates them on later periods, preserving the chronological structure of the market and testing whether a rule holds across different regimes.

Sensitivity, made visible

If a result only works at one exact threshold or lookback, that is a warning sign. Alpha Lab shows you the neighbourhood, not just the peak, so you can distinguish robust relationships from fragile optimizations.

Independent out-of-sample testing

The most recent 20% of history is held back entirely from the search and used only for independent evaluation. Once a rule is catalogued, live tracking continues on newly arriving data that could not have influenced its selection.

Figures captured from a single live run on 18 August 2026 and shown as an illustration of the evidence Alpha Lab provides. Alpha Lab is a research tool: its outputs are based on historical data and are not investment advice or a guarantee of future performance.

Beyond Discovery

From Discovery to a Living Catalogue.

Not every session needs to start from zero, and a rule that worked is not the same as a rule that still works.

Curated

Catalogue

Strategies already identified and saved by the Glassnode team. Browse by asset, direction and risk profile.

Yours

My Catalogue

Save what matters to your team and build a research universe of your own.

Live

Market Pulse

Which catalogued long and short rules are firing right now, in one compact read.

Over time a saved set becomes more than a list of backtests. Alpha Lab surfaces rules whose recent behaviour has weakened, flags overlapping strategies and shows where your coverage has gaps. The question moves from did this work? to is it still behaving as expected?

Private Beta

Find Your Edge.

Alpha Lab is open to a limited group of clients while we shape what comes next. We are looking for research teams and investors who want to test it against real workflows.

Access
Private beta
Metrics
1,000+
Modes
Easy · Advanced · Agents
Built for
Research teams