Add button link to tracker (https://stockfish.dev/chart/) on the test page

[DanSamek]

Makes it easier to access the test tracker directly from the test page.

[Copilot]

Pull request overview

Adds a direct link from the test view page to the external statistical chart tracker, improving discoverability for users who want to inspect charts for a specific test.

Changes:

• Add a “Statistical charts” button on the test view page (hidden for SPSA tests).
• Add a Font Awesome icon for the new button via CSS.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
server/fishtest/templates/tests_view.mak	Renders a new external-link button to the tracker for non-SPSA runs.
server/fishtest/static/css/application.css	Extends the icon pseudo-element rules and adds an icon for the new button.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[ppigazzini]

I have few doubts:

• I don't know (my ignorance) who maintains the tracker, when it started collecting data (it seems mid-November 2025), and what is the retention policy of the tracked data (keep forever? only last N months?)
• The chart of tracked data is mostly a random walk for a SPRT test, a line for a NumGames test. I have the impression that a button with that label overstates the meaning provided by the chart (we have a page with proper statistics for any test). I wonder if simply adding a link in the Live Elo page would be a better solution, but this add a click of course

I call for opinions on this PR :)

[DanSamek]

I have the impression that a button with that label overstates the meaning provided by the chart

I agree, maybe better text would be "Patch tracker",...

But I'll wait on another opinions about this PR.

[vondele]

I have also a bit my reservations on adding too many external dependencies. However, I think it would be interesting if fishtest started to collect some 'time based metrics' (e.g. number of cores/workers/nps over time), or per test some data.

[vdbergh]

The thing is that all information about a test is contained in its current state. The trajectory that lead to this state contains no information.

One says that the pentanomial frequencies are a sufficient statistic for the pentanomial probabilities.

Formally: the probability of a certain trajectory before time t, conditioned on the state at time t, is independent of the pentanomial probablities.