Explore Gets Personal, Stockfish Gets Faster, Skills Get Tested
009 - Customisation features, engine optimisations, and a closer look at how the skill tree formulas are shaping up.
Hey folks, welcome back đđż
I must confess that while I have been making progress on the skill tree (I promise!), Iâve also been a little⌠distracted. Ever since launching Explore Mode a few weeks ago, Iâve been constantly tinkering with it. Itâs the first thing new testers touch, so I want it to feel smooth, personal, and fun from the moment they log in.
Hereâs what I managed to ship this week:
đď¸ Explore Mode Customisation
Explore Mode now feels a lot more like your workspace. I added a bunch of new controls that let you set things up exactly how you like:
Smart Responsive Sizing â the board now adapts automatically to your screen.
Manual Size Slider â prefer precision? Pick a size between 300â700px and watch it change in real time.
Arrow Colour Picker â eight colours to choose from, so your analysis lines can match your vibe.
Chessboard Theme Editor â play around with the look and feel of your analysis board.
Itâs a small thing, but I love how this transforms Explore into something personalised. And once you set your preferences, they stick between sessions â set it up once and youâre good to go.
⥠Stockfish Wrapper Upgrade
Under the hood, I gave Stockfish some love. The Python wrapper has been upgraded, the engine pool expanded, and caching fine-tuned. This means the analysis engineâŚ
Starts faster,
Streams results as they come in,
Feels way more stable under heavier loads.
For testers, this means snappier analysis and fewer âhangs.â For me, it means fewer late-night debugging sessions wondering why brilliant moves disappeared or openings werenât detected correctly.
đ Pen Testing Foundations
Not glamorous, but important. I put together a penetration testing suite that lets me stress-test every layer of the app â frontend, backend, AI engine, even container security. One command, full report.
The goal here is simple: as Rookify grows, I want it to be both innovative and resilient. No fun if the coaching is brilliant but a bad actor can break the doors down.
đł Skill Tree Progress
And yes, I did make progress on the skill tree this week!
A big part of my focus was running the skill leaf formulas through the acceptance testing system I built. These tests are designed to answer one simple but critical question: does the scoring make sense in realistic scenarios?
The process works like this:
For each skill (say, develop pieces quickly or avoid blunders), I run a set of mock PGN scenarios through the scoring engine.
Each scenario defines an expected outcome: good play should land in a high score range, poor play in a low one, with reliability checks to make sure the result isnât noisy.
If a skill passes, it means the formula is behaving as intended â rewarding good play, penalising mistakes, and scaling fairly.
If it fails, it usually means the ranges are off or the conditions are too strict/lenient.
Hereâs where the iterative part comes in: I tweak the YAML formula, rerun the tests, and watch the pass rate climb. Each run gives me a clearer signal about whether the logic is fair and consistent across Elo tiers.
Itâs a grind, but a valuable one. When a skill passes consistently, it tells me the formula is trustworthy and when dozens of skills pass together, I know the skill tree as a whole is starting to feel like a reliable assessment system rather than a fragile experiment.
As mundane as they may look, these skill formulas will be the DNA of the coaching system. Theyâre what turn raw chess events into insights that actually mean something. Too harsh, and players feel punished. Too soft, and the feedback isnât credible.
𧪠Try it Out
The Explore & Practice Modes are still live at:
đđż rookify.io/app/explore
Iâve capped access at 100 testers, and there are still 50+ spots left. If youâve been curious, nowâs a great time to hop in and play around.
đŽ Whatâs Next
Looking ahead to next week, hereâs whatâs on my plate:
Finish Skill Tree Acceptance Testing â I want to push through the remaining skills and aim for a 100% pass rate across the board. The goal is to get the formulas stable enough to start trusting the tree in real scenarios.
Run the Skill Tree on Real Games â so far, testing has been with synthetic PGNs. Next step is pulling in real Chess.com games across Elo ranges and seeing how well the skills hold up in the wild. This will be the real stress test.
Practice Mode Analytics â Iâm planning to add a feature that tracks your move accuracy over time when using Practice Mode. This will give players a clearer sense of whether theyâre improving from session to session.
If all goes to plan, next week will be a mix of validation and visibility with making sure the skill tree works as intended and giving players their first glimpse of personal improvement data.
See you next week,
â Anthon
Chief Vibes Officer @ Rookify
By the way, if youâre a chess player, Iâd love to hear about your own improvement journey.
Whatâs been frustrating? Whatâs actually helped? And what kind of innovations do you wish existed in the chess world? If youâve got 3â5 minutes to spare, please fill out this short surveyâit would mean a lot and will directly help shape how Rookify evolves.
Thanks again