The Main Value Proposition of Math Academy
Want to get notified about new posts? Join the mailing list and follow on X/Twitter.
The science of learning has identified many cognitive learning strategies that can be used to help people learn better and faster. These strategies have been known for many decades in the research literature, some even over a century old, but until now nobody has successfully unified them, much less at scale. The main value proposition of Math Academy is that we developed the tech to unify all those findings together into a learning experience that adapts to each individual student, at scale.
The closest thing to it used to be expensive training sessions with the most serious individual performance coaches, who would typically end up converging on training methods that leverage many of these strategies. However, the key limiting factor was how much manual effort it takes to track all the variables at play and make training decisions – what are the thousands of individual skills of your domain, what needs to be learned before what, how strong and stable is each one of your students on each of these skills, what are they ready to learn next, what skills are fading and need refreshing, what new skills implicitly refresh old ones as component subskills, what is the most efficient set of tasks they should do today to acquire new skills while maintaining old ones, how much practice do they need on each new skill before they’re ready to move on, and so on.
Even if you’re an expert individual performance coach (e.g., elite math tutor) who has a good idea of what efficient training looks like, only has a few students to train, and is training each student in their own individual sessions, the sheer amount of bookkeeping and computational optimization required is so overwhelming that you just can’t do it all manually. You can try to put together some spreadsheets and construct a learning experience that is directionally correct, but at the end of the day your human limitations put a hard limit on how high you can push the level of learning efficiency for your students.
And that’s the most optimistic scenario. If you’re a teacher with 6 classes to manage and 30+ students in each one… forget about it. It’s hard enough to keep all these students learning at all, much less make it efficient. What do you do when each student comes into your class missing a bunch of prerequisites for what you need to teach, and they’re all missing different things? Without technology to deliver an individualized learning experience to each student, it’s an intractable situation. Even if by divine intervention, a practice problem magically appeared that the top student didn’t already know, the bottom student was actually prepared for, and that would actually push forward in the course (not just a supplemental enrichment activity)… what do you do when the top student finishes in 45 seconds, gets it right, and is ready to move on, while the bottom student makes a mistake 3 minutes in and needs more practice? Even if that miracle were to happen, minutes later you’d be back in the same intractable situation.
That’s really the problem we’re solving – not just identifying the ingredients of efficient learning, which is the easy part, but building the actual technology to pull it all together. That’s the hard part, and it has required us to build one of the world’s most sophisticated expert systems, which includes not only a massive corpus of top-quality educational content broken down into bite-size pieces connected up in a knowledge graph, but also novel algorithms such as Fractional Implicit Repetition and Spaced Repetition Compression, which enable us to overlay a student’s activity on top of that knowledge graph, figure out their mastery level on every single topic, and serve them new learning tasks that maximize their learning per unit time.
And the end result of building that system is that students who put forth honest, consistent effort develop mastery several times faster than they otherwise would in traditional classrooms – mastering a grade level in several months, mastering advanced math at shockingly young ages – with that mastery transferring to external assessments outside our system (e.g., MAP scores, AP exams).
Want to get notified about new posts? Join the mailing list and follow on X/Twitter.