What it is

In the Workspace, a node is a step on your Workspace canvas. Nodes are connected into a sequence so you can test how different policy changes interact.

The result is a visual work trail you can retrace: it shows what your next step is, and how the connected inputs turn into a single workspace backtest run.

This article explains the node “building blocks” in user terms:

  • what nodes represent,
  • how the sequence changes what gets evaluated,
  • how executing accounts affect results,
  • and how to interpret policy inputs that come from different timeframes.

What you can do

  • Understand the purpose of each node before you connect it
  • Build a node sequence that reflects your intended sequence
  • Choose executing accounts so you can compare outcomes across different account contexts
  • Combine different policy changes (entry and exit edits) in the same workspace
  • Mix inputs built from different report timeframes, and label your results correctly

How it works

Nodes as canvas steps

Think of each node as a step that either:

  • establishes a starting cohort / context,
  • applies a policy change, or
  • produces the workspace output you want to review.

In practice, the flow is:

  1. Pick the report / simulation scope you want to work from.
  2. Run a backtest on that scope.
  3. Save the result as a policy.
  4. In Workspace, place one or more saved policies into your sequence.
  5. Optionally combine those policies into a combined policy.
  6. Run a new backtest from that connected workspace setup.

Workspace gives you a clear visual way to organize saved policies, combine them where needed, and then run the next backtest from that connected setup.

Sequence matters

The sequence is the visual order of your work. It helps you understand:

  • what came first,
  • what was created from it,
  • and what the current backtest is based on.

As your strategy work grows, this order makes it much easier to retrace what you tested and understand how one report, policy, or backtest led to the next.

For combined policy steps, lineage is preserved from the source policies you selected. This keeps the chain understandable when you revisit the canvas later.

Executing accounts: who runs the policy

In Workspace, executing accounts are the account context you assign to the strategy and policy versions you are using in that sequence.

This is mainly for clarity and retracing:

  • You can easily remember which account context your results belong to
  • You can compare outcomes for different policy setups (or different account contexts) without losing your work trail

If you start a new sequence, pick your executing account first-so your new workspace starts from the context that feels comfortable and familiar.

Policy inputs inside nodes

Workspace nodes are connected using policy versions created in reports and simulator tools.

When you connect policy inputs, the key question is: what kind of change is this step making?

Common policy input types include:

  • Entry filters: include/exclude trades based on market context at entry.
  • Exit management rules: apply stop-loss / take-profit distance changes to manage exits.
  • Combined / bundled policy inputs: combine different policy changes into one step so the workspace evaluates them together.

Mixing policies from different candle timeframes

It’s common to connect policies that were built using different candle granularities (for example: 1m, 5m, and 30m) into one workspace sequence.

This is useful because it lets you test policy combinations across multiple candle resolutions in one run.

How it works:

  1. Tessera uses the most granular candle timeframe as the lock for the combined run.
  2. On each trade, each connected policy segment is still evaluated on its own configured candle timeframe.
  3. The combined result shows how those cross-timeframe policy combinations perform together.

This allows you to compare whether a given combination leads to more or less favorable outcomes.

Limitations

  • Workspace results are counterfactual: they explain what would have happened on the stored closed-trade cohorts behind the connected inputs.
  • Workspace does not model live execution bar-by-bar or tick-by-tick fills.
  • Mixing timeframes is “valid as a test,” but it’s not guaranteed to match intuition - always interpret results with the cohort scope in mind.
  • If coverage is weak for any underlying input, labels and interpretation can become less reliable.