.png)
1 day ago
4
I feel personally connected to his experience. Two months ago, I started trying Math Academy (I'll call it MA). I found it to be a serious tool for learning. It reminded me of my school days when I used to study and do math problems. Math was always a subject I was good at. But after I started working, my interests changed, and I slowly stopped putting effort into math. Suddenly, I realized the last time I systematically learned math was actually five years ago. Because MA helped me want to learn math systematically again, and it really gave me a good experience, I wanted to recommend it to more people.
But because my own math skills are already quite good, I was worried that my experience might be very different from others. So, I remembered a friend who sent us an article before. He dropped out of middle school and hadn't learned math systematically since then. But he was very interested in math and was always looking for a good way to learn, which made him quite frustrated. I recommended MA to him and hoped he would try it for a month. I wanted him to share what MA looked like from his point of view, and how he felt about it.
The article below is over ten thousand words long, but every sentence comes from his own direct experience. As I read it, I could clearly see the positive effect MA had on him. This made me feel even more sure that technology can greatly improve education. I also hope that this article can make all of you feel the same excitement I felt when I read it.
June 1, 2025
Jarrett Ye
The original article in Chinese: 【读者来稿】我与 Math Academy 的一个月
This is the story of my experience, and I've divided it into two parts. The first part is a short version, about three thousand words. A large language model helped me write it based on my original notes, and I checked and changed it. The second part is my original notes. I wrote these notes whenever I had new thoughts during the month. So, this part is long, about nine thousand words. If you want a quick read, you can just read the first part. If you want to know more details, you can also read the original notes. Okay, let's start.
My relationship with math has always been hard to explain clearly. Saying I have a "weak foundation" is not strong enough – my math skills are mostly at an elementary school level. I only understand a few ideas here and there, so my understanding is not complete. For me, normal ways of teaching math were very scary, making it impossible for me to learn well. But do I really hate math? Not really. Whenever I solve a problem, it always makes me feel a special kind of excitement. This shows I still like this kind of mind game. As I learned, I slowly understood that my fear of math – sometimes it was almost like I hated it – was all because I lacked basic knowledge. But to fix this lack of basic knowledge, normal ways didn't work, and hiring a tutor was too expensive. These hard problems were like a wall I couldn't get over, not like fun mind games. It seemed like math had closed its doors to me forever.
On April 30th, I started using Math Academy (I'll call it MA), and it was a surprisingly good experience. Its learning system, which changes based on what I need, showed results right away. It wasn't perfect, but it helped put my broken knowledge together and gave me a basic understanding. On the very first day, I felt that fun, game-like feeling again, which I hadn't felt in a long time. The green feedback that appeared when I answered questions correctly was very satisfying.
I quickly noticed the system could adjust itself. When I made a mistake, it seemed to give me new questions about that mistake. If I made too many mistakes, the questions got easier. If I answered some correctly, the questions got harder again. For example, after I got the first two questions completely wrong, the questions became much easier. After I solved these easier questions, the system gave me two more questions. It was like starting over, but the hints were clearer – they directly showed the hidden way to solve the first problems. This system of changing the difficulty based on how I did is very interesting.
However, on May 1st, I found a small problem: if I kept making mistakes, the lesson would end (I guess to go back to an easier level), but the basic teaching didn't follow right away. So, the system that changes how it works together is not perfect. MA does suggest reviewing related ideas using "small links" (short, not official lessons), which have explanations and exercises. But honestly, this information (for me) should have been part of the main course. I didn't feel it was a review at all; it was completely new knowledge! Even when I clicked and learned it (which was like learning it for the first time), the system called it "review" and gave no experience points (XP). That felt really bad. But later I found out this wasn't a big problem with the system; it was because I had set the course difficulty too high myself. About the course itself, the collection of questions is very complete, it covers things well, and it makes learning comfortable. Even for the same course, if I didn't pass it before, the questions would be different when I tried again. Sometimes it even has "traps" – you think you need to use a new method, but actually, you solve it with old knowledge, which is interesting.
From the beginning, I wished I could tell the system which ideas I didn't know. This was especially clear on May 2nd, when I first learned about logarithms. I found I only understood 30-40% of it. At that time, I basically knew the other ideas in that course and could continue by skipping logarithms. But because there was no way to tell it directly, to "trick" the system into suggesting a logarithm course, my only way was to purposely answer logarithm questions wrong (I thought this was the only way at the time). This was actually a waste of time – too many mistakes would end the course right away, and I couldn't learn the next things. It felt like I lost more than I gained. I couldn't even find a way to "quit course"; if I left and came back, I had to continue. So, I could only get rid of the course by purposely answering incorrectly. I always felt that having an "I don't know this idea" button could give more exact control. It would be like giving feedback, helping the system quickly give new lessons I needed. Later I learned that its way of adjusting is made less sensitive on purpose, mainly to stop people from trying to cheat the system with unusual short-term actions. But this clearly needs to be improved – what can a user do if they really want to learn a specific small topic?
By May 3rd, I really felt a "strong pull," almost like I was "addicted," though I wasn't completely lost in it. I started to care about the XP leaderboard. This reminded me of some criticisms of MA – that its XP system might make learning like Duolingo, changing it into just getting points. Even though I knew this could be a problem, for me, the game-like rewards are currently better than the bad parts.
I looked closely at some reviews. People who commented, like Michael Pershan and Dan Meyer, say that MA focuses more on the steps to solve math problems rather than helping students understand ideas deeply, like a good tutor would teach. Pershan even said he felt like a student in a "cram school" (a special school to help pass tests), memorizing formulas for exams but not really understanding them. However, he also said that MA made it very easy for him to remember things.
I understand what they mean – MA might not be the best tool to make future mathematicians. But for someone like me, who needs math as a tool to use in other areas (I am interested in complex economics and social network analysis, meaning I need to learn graph theory eventually), MA is a very helpful thing. Without it, I don't think I would have become interested in math again. The worry because I needed math but couldn't learn it had always bothered me, and MA showed me a way I could go.
But how can I explain my general opinion? It's really hard to agree with how they criticized it. MA uses complete facts from how people learn to support its teaching methods, while critics seem to use their own experiences more. For a subject like math, which needs a lot of practice and remembering steps (which MA calls "automation"), can you really get better just by "understanding"? If I don't even know basic algebra, how can I learn calculus? It would be too much to think about. Frank Hecker's summary is something I really agree with: "Learning doesn't need to be hard work – in fact, hard work can have the opposite effect... Being able to do things automatically is very important." He also noticed that some math groups have a kind of "test" you must pass, saying that you must suffer first to become good at math. I am very glad MA has avoided this idea. Learning should not be painful. It can be hard, but it shouldn't be very painful.
My goal is to use math as a tool. I (currently) don't want to understand math very deeply, so MA's way is very good for me. I also agree with Hecker about controlling how fast I learn – "less is more." Getting a little bit, 30-50 XP per day, feels like I can keep it up better than trying hard to get hundreds of XP.
The first timed test on May 4th was a clear warning: I was too slow. Each question took me 2-3 minutes, much more than the 14 minutes allowed for 8 questions. Then on May 7th, the way the system adjusted felt a bit strange. It suddenly gave me a lot of trigonometry, but I had completely forgotten the basic ideas of the unit circle. Seeing tan, cos, sin gave me a headache. This was an old problem again – unless I purposely answered wrong, I had no way to tell the system I needed to go back to basics and relearn (I really needed a basic trigonometry course), and purposely answering wrong made me lose XP. This experience was very frustrating, especially after many correct answers, only to fail completely because of a few mistakes on stuff that was too hard for me. However, that day, the attraction of the leaderboard also showed itself – the hope of getting back to number one kept me going for four and a half hours, and I earned 76 XP. I even sent a message to customer service about it.
On May 8th, I found a better way:
Change course goals to something easier: If I chose calculus but had trouble with basic algebra, I should choose an easier course as my goal.
Turn on "Holistic Mode": This would start a full test (maybe long, like eighty questions in three hours) to fix all the things I didn't know, even if they weren't things I needed to know for my current goal.
This was exactly what I needed! MA's Q&A explained that the system is slow to react to mistakes on purpose to stop cheating. For real gaps in what I knew, especially for someone like me whose basic knowledge was mixed up, changing the course goal and taking the first test again works better. This helps the system quickly update what it thinks my level is. The official reason for no "I don't know" button (people would misuse it) also made sense – now there's a better way to give feedback by changing course goals.
On May 9th, after changing to an easier course, the experience indeed got much better. My estimated time to finish Algebra II was moved later, from July to October, likely because the 'Holistic Mode' added more basic stuff. Things became clear: high-level first tests don't go deep into very basic ideas. If I chose "Calculus" but had trouble with a "basic" idea for calculus, the system might only suggest more things at that same level, not knowing my real problem was with earlier stuff. Changing to an easier course tested me again more accurately. MA is like a new tutor, learning more about me each time we work together.
Around May 10th, I started getting XP faster per minute, reaching about 60 XP per hour. This showed that when what I knew matched what I was learning, I was getting better at doing basic things automatically. On May 14th, I found myself lost in studying for over four hours, and the leaderboard definitely made me want to "do one more lesson." I got much faster at solving problems, and I even started choosing MA instead of Anki to learn new words – using MA was more fun, not like a boring job.
A big step happened on May 18th: I finished a timed test on time with all answers correct for the first time! My XP per minute seemed to stay around 0.76-0.84, which was almost the same as what they said: "1 XP = 1 minute of focused time." So I was focused about 75%-85% of the time.
Around May 22nd, I started to feel tired of it, which often happens to me when I start new habits. But the idea of keeping my place on the leaderboard (on May 24th) made me want to catch up, and I quickly got 150 XP in one day. Now I finish quizzes much faster, often with half the time left.
Then on May 25th, I had a sudden understanding. While learning the area of an equilateral triangle (a triangle with all sides equal), I really felt that math was beautiful. The formula that looked strange (√3/4 times the side length squared), when I looked at it using trigonometry (sin 60°), its logic suddenly became very clear and perfect. It nicely connected basic geometry (base times height divided by two), trigonometry, and the Pythagorean theorem, all giving the same answer. This experience made me agree more with what MA believes: you have to learn basic ideas first to see how beautiful math is. When math is just a lot of symbols I can't understand, you can't see any beauty.
Very quickly, it was May 31st, a whole month. Even though stress from life and work made me tired of learning, I changed my goal from 50 XP a day to making sure I got 350 XP a week, often studying a lot on weekends to catch up. From what we know about how people learn, this might not be the best way (learning a little bit often is better), but it's very important to keep going in some way.
Looking back at this month, I was surprised by what I learned: unit circle, trigonometry, basic geometry with triangles, logarithms, factoring, complex square roots, inverse functions, complex numbers, solving inequalities... This stuff is very basic, probably not harder than middle school math, but for me, it was a big step forward. I had tried normal textbooks and online courses many times, but never stuck with them for more than a week. MA helped me study for a whole month without stopping, not only without feeling like I didn't want to (being tired from life stress doesn't count), but I even plan to pay for another month.
MA's design, based on science, worked for me. Old ways of learning didn't work for me or even made things worse. I thought about a tutor, but it was too expensive. This once made me feel hopeless, thinking I would never learn math. Now, I've finally found the only way I can learn. Not just a second month, I think I will keep using it for a year. I used to think math was like a very high mountain that normal people couldn't climb, but now I want to work hard and learn everything in MA. Yes, I've only learned basic stuff now, and maybe one day I won't be able to understand more abstract ideas. But for sure, with what I know now, I'm still far from how much I can learn.
I really thank the people who made MA. They didn't just give me knowledge, but also something very valuable – belief that I can learn. This is very important if you want to keep learning for a long time. It's true that Math Academy gave me back my confidence to learn math. Just this one thing is a gain so big you can't measure it.
I really felt its system that changes for me had some effect. My math basics are very poor; you could say I'm mostly at an elementary school level, but it's up and down. In this situation, the system, even if not the best, helped me get a general idea.
I've always been afraid of math, but when I solve problems I can do, I often feel great. This shows I really like this kind of mind game. But why am I afraid, or sometimes even feel like I "hate" math? I think it's only because I lack so much basic knowledge. When I see problems I don't understand at all, I don't feel like it's a game, only confusion because I can't understand.
The first day using MA was quite nice. I really got back that game-like feeling I missed. At least learning was enjoyable, and the green feedback when I solved a problem was very clear.
If I make mistakes on problems, it seems to give new problems based on the mistakes. If I make too many mistakes, it gets easier, and if I get them right, it gets harder again. I clearly saw that if I got questions wrong, there would be more questions, and it happened step-by-step: 2 questions to try, all wrong made it easier; getting the easier questions right made it harder and gave 2 more questions. These new questions had very clear hints, the solution was almost shown (this solution was for the first 2 questions, but it wasn't clear from how the first two questions were written). This wasn't there in the first 2 questions. This shows it has a difficulty level that changes based on whether the user is right or wrong, which is quite interesting.
If I keep making mistakes, the course will end; I guess the difficulty will go back down.
The part that's not perfect is probably that the system for adjusting things together is a bit incomplete. For suggested related ideas, because there are some I really don't know, it treats them as review. This means I don't get XP when I do those older things. Of course, overall, this is just something that could be improved, not a big problem.
The way it gives questions is quite interesting; sometimes it sets traps, making you think you need to use the new method you just learned, but it's actually an old method.
At first, I wondered if users could actively choose ideas they want to learn, since computer programs might not cover everything. But later I learned that the whole testing system changes automatically, so I don't need to actively review when I don't know something. Thinking about it, this is probably the best way. If the test decides based on whether the user makes too many mistakes, then if I click the review links it gives, the system might get the wrong idea. It might think I've already learned that idea, but actually, I'm learning it for the first time. So the best way to give feedback is still to do the practice problems it gives me. Making many mistakes shows I haven't learned the basics, and the system will naturally adjust and suggest more basic material. (1)
Today was a bit better, but there's one thing to watch out for: when you see complex formulas, choose your answers carefully. It's easy to misread, and if you click the wrong thing by mistake, you can't fix it, which is awkward.
I found that my idea (1) might still be a bit off. Today I learned logarithms, but I quickly realized I was in a difficult situation. Even though I do know some of the material, it might only be 30-40%. To learn this, I would need to give up the whole course, which feels like I would lose more than I gain. I couldn't even find an option to quit the course, so I could only pretend I chose wrong. Would it be better to add an "I don't know" option for each question (or even the whole course)? This would let us adjust the system more precisely, to learn the specific missing pieces of knowledge.
I clearly noticed I was starting to feel something like "addiction." It's not like I was completely lost in it for a long time, but I can feel that "can't stop" feeling. And I clearly care about the XP ranking. This reminded me of what some people criticize about MA: "Will the XP design make MA like Duolingo, or even change learning into just getting XP?" Maybe this is partly true, but I think as a good game-like reward, so far, for me, the good parts are more than the bad parts.
Here are some criticisms of MA:
A balanced review of Math Academy[1]
Math Academy Wants To Supercharge Your Learning[2]
It Is Fun to Pretend That Hard Things Are Easy![3]
Reading what the critics wrote, they have some good points. For example, they say MA's training methods focus more on the steps to solve math problems rather than helping students think like mathematicians – or understand math deeply – like a real tutor would. But I think this depends on the person. Maybe you can't get a deeper understanding of math with MA's way (introduction + example + solving problems). Or maybe MA is not good for making mathematicians, but for people in other fields who use "math as a tool," MA is very much needed to help them start learning math. From my own use these past few days, if I didn't have MA, I'm afraid it would be very hard for me to become interested in math again. But my future work and studies absolutely need math (I'm interested in complex economics and social network analysis, and will probably need to learn up to graph theory). This worry has truly been bothering me – but MA has shown me a way to reach my own goals.
Also, I somewhat disagree with what the critics say. At least MA has shown complete facts from how people learn as proof. The critics seem to just ask questions based on their own experiences, which is clearly not strong enough. From what we know about how people learn, for a subject like math that needs a lot of practice and remembering steps, can you really move forward just by "understanding"? A lot of boring, step-by-step, or, as MA says, "automatic," things you do without thinking are also very important. If I don't even know basic algebra, can I start learning calculus? That would be too much to think about.
A lot of calculation and boring practice indeed make math lovers feel that "math is more than just exercises." Yes, that's right. I've memorized the formulas, I know when to use them, but I haven't deeply understood these formulas from a math point of view, as Michael Pershan said:
While working at MA, I felt like I became that kind of cram school student. Spaced repetition does work—I was indeed able to remember the formulas for variance and expected value long enough to ace quizzes. But I felt my understanding was weak, and honestly, it frustrated me. I still don't know what a geometric distribution is. I could remember it because their targeting was very precise. Their probability course isn't made of units, but of 180 topics, each of which itself contains 2-3 skills. The problems I practiced were exactly the same as the examples, just with different numbers. Reviews and assessments were the same—I never encountered any problem in practice that I hadn't seen an explanation for before.
However, even Pershan himself (Posts [4]), very honestly admitted that MA made him remember things very well, even "extremely smoothly" – though he personally didn't like this.
I personally agree more with MA's ideas, as Frank Hecker summarized:
Learning doesn't need to be hard work—in fact, hard work has the opposite effect. "The way to improve students' ability to make mental leaps is not to make them jump further, but to have them build bridges so they can leap from the bridge." Being able to do things automatically is important, and Math Academy needs to check this. You don't need complete automation to move to higher-level topics, but a lack of automation will eventually stop students from progressing.
At least in my personal use, this form of MA is something I like very much, and it fits my purpose well (I'm an active learner and I use math as a tool). To continue with Hecker's words:
Problems are usually variations on the same theme—sometimes very small variations. Given this, I can understand why some people would hate this teaching style. Some students seem to have a tough-guy attitude, believing that learning mathematics is not supposed to be easy, and one must first complete exercises in textbooks like "Baby Rudin" and "Papa Rudin" to be considered good at math. It's like a bullying ritual, where each new group must suffer for the previous group to feel their own suffering was justified and worthwhile. The creators of Math Academy clearly do not share this attitude, and I personally am very glad about that.
Math should not be painful, or more generally, learning should not be painful. Pain is different from difficulty. When I find something hard to do but enjoy it, that's difficulty. When I see something and it makes me feel a lot of bad emotions, that's pain. Even though MA emphasizes speed (4 times faster than old methods), my views are similar to Hecker's, mostly on these two points:
You need to actively control your learning speed to get a "less is more" effect. You don't need to get hundreds of XP every day; staying at 30-50 is enough. This is also what I've learned from many years of studying – sometimes it's just like that, you slowly learn something without knowing when, and that's the result of time and effort building up.
Pershan and Meyer's criticisms from the point of view of traditional textbook teaching (which focuses on proofs) are not wrong. But for me, especially because I learn with a clear tool-like purpose, I'm not very interested in math itself, and I don't like abstract and proof-focused learning methods – MA is excellent.
Today was the first test. I clearly found my problem-solving speed was very slow; twice in a row, I didn't finish before it ended. From an "automation" point of view (doing things automatically), there are indeed some areas that still need practice. Looking at the timing afterwards, I found that each question took an average of 2-3 minutes. If you calculate it that way, 8 questions would take about 20 minutes, but the test is only 14 minutes long, which is too short.
Today, I clearly found the system that adjusts itself still has some problems. It suggested a lot of trigonometry material to me, but I had completely forgotten the basic knowledge of the unit circle. Seeing tan, cos, sin gave me a headache, and then the test even had questions asking me to calculate tan, which I couldn't do at all. I found that apart from purposely getting questions wrong, there was no way to tell it if I could go back to basic trigonometry knowledge and start over. And purposely getting too many questions wrong would take away my XP, which was really uncomfortable. This experience was indeed not good. I still think adding a way for users to give feedback would make it more complete. For example, since it shows the detailed list of topics, I could actively choose the content where it guessed wrong about what I knew, and tell it that I don't know these parts at all. I'm not saying it should rely completely on user feedback, but making the user's active input one of the things the system considers, at least to affect how sure it is, I think would be better than the current way that relies entirely on getting questions wrong. Indeed, for users who "don't know what they don't know," guessing based on their actions is most effective because user choices might be completely wrong. But what about users who "know what they don't know"? This is clearly a blind spot; the system hasn't left space for these users.
In short, today was very frustrating. I clearly found a lot of unknown things I hadn't learned, which made the course it gave me a bit too hard, and learning was very uncomfortable. Especially when I got all the first dozen or so questions right, but then at the end, there was material that was too hard, and three mistakes in a row made me fail everything. All the previous correct answers were wasted, not to mention the wasted time, and my XP even went down. This experience was truly awful.
Today I really felt "addicted," especially with the leaderboard. When I was in second place, I wanted to keep doing one more lesson to become number one. The game-like strategy is very effective. However, today it took me four and a half hours to get 76 XP. Even though I made a lot of mistakes in between, it clearly doesn't match the official rule of 1 XP = 1 min. Even my regular speed is 1.5-2 minutes per XP. I don't know if the officials change how much XP you get based on speed. I sent an email to the officials asking if there's any way to solve the problem of it not suggesting more basic courses, and I hope to get a reply.
Today I discovered the correct way to use it. I had been complaining about not having a way to give feedback, but strictly speaking, you can actually give feedback yourself. The most important rule is, don't try to do too much at once. When you find that your level is too far from what the system has guessed, you need to:
Go back in the course. For example, if I chose calculus but have many problems with basic algebra, I should clearly aim for an easier course instead of directly choosing calculus.
Turn on "Holistic Mode." This might give me a complete extra test lasting up to three hours with eighty questions, to help me fill in the missing knowledge (even if this knowledge is not directly needed for my current course). This mode is very effective for building a foundation, but the downside is that it will take longer to finish my goal because various basic courses will be added.
You can do both of these at the same time. If my goal is to learn math better – not just to quickly learn one course and be done with it – then going back while turning on holistic mode might be better. Making your foundation strong is never a bad thing. Even if I learn math just as a "tool," more practice will also make it easier for me to learn advanced math later.
According to the official Q&A, lowering your goal is a better method than purposely getting questions wrong. This is because there are measures to prevent cheating that stop students from easily moving from the learning zone to the easy zone. So, the system for giving feedback on mistakes is purposely set to be quite slow. This is clearly a good thing for students who want to easily get XP, but it depends on the system correctly measuring the student's knowledge. For example, if I get some right and some wrong, then a single mistake doesn't strongly show that I lack knowledge. Especially when my knowledge is mixed up, the system will need more time to judge. If my wrong answers are because I really lack knowledge and I purposely want to go back, the system, based on my pattern of getting some right and some wrong, might think I'm cheating. So, it would more likely make me redo things rather than just go back.
So it's obvious, when I'm in such a situation, it means I should choose a new course goal and take the first test again – and indeed, after I chose a lower-level course and finished the first test, the system immediately updated my knowledge level. I did go back to what I thought was a better level. At this point, both for me and the system, we were much more sure about my knowledge. It seems like you only get these first tests when you change goals; changing the goal itself is a chance to take more tests, which is also a benefit of going back.
I continued reading the Q&A. The official reason for not suggesting to turn on "Holistic Mode" is that it might include too many basic courses that are related but not important for the learning goal, and this can delay finishing the learning goal. Indeed, compared to always adding basics, only adding the necessary knowledge related to the goal is a better reward; otherwise, the time to finish the goal will be too long. If my aim is to reach the highest level, then the total time will actually be the same, because the basics always need to be learned. However, learning necessary things first will make the goal clearer and motivate people more. This makes sense. I plan to try holistic mode for a while to see how it works.
About why they don't add an "I don't know" feedback button, the official reason is that it's too easy for students to get into a habit of clicking it. Whether students click it on purpose, or because they don't have enough information, clicking "I don't know" whenever they find something hard is clearly more bad than good. Okay, I can't deny this reason. Considering MA is not just for adult learners (even adult learners who lack confidence might misuse this button), maybe not adding more buttons is reasonable.
Of course, before today I might not have thought so, but today I found "going back to an easier goal" as a way to give active feedback. After redoing the test, the system also became more accurate. Now my need for a button is indeed not so strong. Not having a button certainly relies on a good system; the certainty from a first test done only once is indeed not high enough.
Today I haven't had the big test for holistic mode yet; it might come out slowly later. But there's another possibility: based on the current assessment, has a hidden holistic mode already started? I noticed my goal (Algebra II) completion time increased from July to October, most likely because more basic knowledge was added.
About the adaptive testing, after looking at the official explanation, I have a clearer understanding of what happened before. Personally, I think this is really a matter of balancing things. Because of the number of questions – as they themselves said – an adaptive test can't ask thousands of questions to fully understand a person's knowledge, so it must use computer programs. So, when I choose a high-level course, the program, for some probability reasons, will definitely not include too many easy problems. Imagine, I choose calculus, but then I get tested on a bunch of addition, subtraction, multiplication, and division, or even more basic stuff. Nobody would like that. So "basic" here must have a relative meaning. Compared to advanced math like calculus, what level is "basic"? According to what I've seen, its "basic" and my "basic" are definitely estimated differently. High-level goal tests naturally have more high-level questions. Even the "basic" questions in them are not what I think of as basic questions. For example, it might guess I lack basics because I answered an "advanced basic" question wrong. But this "advanced" might be a third-level basic, while I am missing first-level basics. Then it will keep suggesting third-level courses to me, but this third-level content all depends on the first level, so of course, I will learn very unhappily. Plus, as I said before, the system has a slow adjustment to prevent cheating or bad performance. This means if the first test didn't go well, later learning will also be very bad. Maybe – to be honest, I'm not sure – after a long enough time, it might also suggest first-level courses to me, but instead of waiting a long time, why don't I just go back to a basic course goal?
After going back today, the experience indeed improved a lot. And as I said yesterday, after I did another basic test, it clearly allowed MA to understand me a bit more. So I can imagine a very natural result: as time goes on and I use it more, MA will understand my knowledge better and better. The courses suggested to me will increasingly fit my "learning zone," and I will learn more and more smoothly. The terrible experience I had when learning calculus before will definitely be much less. After all, when it clearly knows I lack first-level knowledge, there's no reason to keep suggesting third-level content to me. Not to mention I've also turned on holistic mode; when the big test comes later, even the most tricky knowledge gaps can be cleared up as more data is added. In the official words, MA is like a personal tutor. Currently, we've just met, so it's also slowly getting to know the details of me, this student. A tutor who doesn't understand the student well enough certainly can't help fully, and understanding takes time, by getting more data as the student keeps using it.
Of course, I won't say MA is super magical; after all, everything has its limits, and it's no different. But if I compare it to other things, especially traditional classrooms – then I really like this model a lot. I couldn't stand traditional math teaching for even a minute, but with MA so far, I average at least more than an hour a day. At least for me personally, MA is even the only way I can learn math. So no matter what problems it has, as long as they're not too serious, I can probably accept them. After all, learning with it is currently the only way I can feel joy and fun in math, so why not use it? Anyway, my purpose for learning math has always been as a tool. Here, I certainly understand that some people are not happy that MA doesn't teach much about the beauty of math (Andy also complained that the examples are not beautiful or connected enough [5]), and that it's purely calculations and steps – but for people like us who are not interested in math at all and learn it purely to use it, MA is truly a great help. For me, maybe in the future, I will be able to feel the beauty of math, but before that, I have to be able to use it well. Maybe those who truly love math wouldn't stop someone who might appreciate its beauty later from trying to learn it? After all, for me, traditional textbooks have completely failed; without MA, I might not even touch math again. You see, what was impossible before has now become possible. At least from the point of view of letting more people continuously learn math, math lovers probably don't need to see it as a bad thing; it's not a big problem.
Today I found the person in first place on the leaderboard got up to 300 XP; I need at least 80+ to catch up. But never mind, I'm happy with 60 today; I'll let him be number one. For the past few days, it's just been the two of us fighting for first and second. I don't know if he saw that I was leading all day for the previous few days and decided to try really hard. If that's true, then he indeed won (why is there so much competition even in the Bronze league!). However, I have personally felt the power of making it like a game. As long as it's not twisted too much (like some people say, learning just for XP), then within a certain limit, it's a good motivation. I admit that sometimes I also want to do one more lesson, just so the XP from that lesson can get me to the top of the leaderboard. At the same time, I also understand that some people don't like this system, especially since the higher the rank, the harder it is to move up (I can't imagine how much Diamond league people study in a day). But since it can be turned off, everyone can choose what's best for them.
Today I found that the time to get one XP started to speed up. I can get around 60 XP in an hour, which wasn't this fast before. This shows that after the level is matched, the XP you get can match too. If solving problems is too slow, it means the "automation" (doing things automatically) of some basic skills is too low, and you need to learn more basics.
Today I accidentally studied for over four hours again, and still felt like continuing, not quite done. To be honest, I haven't played games for a long time, but this really gives me a game-playing feeling. For example, the leaderboard: after finishing a session, I casually look at the board, just like pressing Tab in an FPS game. Seeing my rank is a bit behind the one above, I always feel it's best to do one more lesson. If I'm not too tired, I'll just study a bit more, and I'm only happy when I reach number one. Although some people really don't like leaderboards, I might really like this kind of thing. However, I also need to be careful not to study too much, otherwise, the higher the rank, the harder it is to move up; a moderate amount is enough. From the point of view of learning and understanding, the most noticeable thing recently is that my problem-solving speed has increased, and I can somewhat keep up with the speed of some quick tests. Although I still make quite a few mistakes, at least it's better than at the very beginning, so I have indeed learned and understood a lot of things.
In the afternoon, I originally planned to memorize Anki cards, but I accidentally "finished all of" MA for the day. I have to say this indeed has more positive feedback than Anki; at least the official design idea has worked well for me. I find my motivation to memorize vocabulary cards has dropped a lot recently, and I often have several days' worth piled up. If today I had unfinished tasks for both MA and Anki and didn't have enough time, I found I would really choose MA over Anki – the reason is simple, MA is more fun to do, and I naturally try to avoid boring vocabulary cards.
Today, finally, I got a perfect score on a test within the time limit for the first time. So it's indeed true, its automation training is also useful. And I feel it adjusts; this time, I clearly felt the answering time was not so tight for me. This shows the questions given really matched what I've been learning recently. Today I studied for about two hours and 20 minutes, and the XP I got per minute was 0.76. I think it has stayed at this speed. In that case, 50 XP per day would be about 65 minutes. But if I take out the 15 minutes in between when I was very distracted chatting with someone while studying, the actual XP per minute is probably 0.84.
The officials say 1 XP is one minute of focused time. So it seems this can also be understood as my focus rate being around 75%-85%. Perhaps this is the per-minute XP count for most people? After all, if you study for a long time, it's impossible to be focused every minute. For two or four hours at a time, you always need to rest in between. So you can imagine that the longer the time, the lower the focus rate might be. Of course, how much XP you get also depends on difficulty. If the problems are too hard and I keep making mistakes, then spending half a day with 0 experience will obviously lower the per-minute XP count. So, it can be used as a general idea of focus rate, but you can't say it's always true.
It seems I've entered a period where I feel tired of it. This is normal for me; anything around 20 days starts to become a bit on and off. Generally, as long as I get past this period, it will become smooth. Overall, I still want to continue using it. I want to try and see if I can do fixed things at fixed times; perhaps this will make it easier to continue using it. At least overall, I think it can still give me the motivation to continue. Currently, I don't feel much like avoiding it; it's just that I've entered a flat period for myself. Once this period is over, the habit can be properly formed.
I should say the leaderboard is still somewhat useful. When I found I might drop a rank, I started to feel motivated to make up for what I missed for the week. Today I quickly got 150 XP in one day, and tomorrow I might be able to get another 100, so this week's 350 will be complete. And I find that I'm doing quizzes much faster now. At least before, I often couldn't finish them, but now I finish and find I still have half the time left. This means that the current system might have found my actual level, and the questions given are quite suitable for my current level.
Today I truly felt the beauty of mathematics from calculating the area of an equilateral triangle (a triangle with all sides equal). All parts fit together perfectly; those formulas that looked tricky are actually logical results. That formula for an equilateral triangle that seemed strange, √3/4 times the side length squared, can actually be understood at a glance if you've learned trigonometric functions. This formula can come from "base times height divided by two," which you learn in elementary school, by just thinking of the height as related to sin 60° on the unit circle, then multiplying by the side length to scale it; then everything becomes clear. Here I had a moment of sudden understanding, seeing how I got the same result from three different ways: geometric area calculation (base times height divided by two, splitting rectangles like we learned in elementary school), using trigonometric functions, and the Pythagorean theorem. Although the knowledge is quite basic, I agree more with what the officials said – to understand math, you first have to learn these basic ideas. Otherwise, no matter how beautiful math is, I simply can't feel it. Especially when math to me is like a secret code, forget beauty, just seeing a bunch of square roots gives me a headache; it's too much to think about. It's necessary to first know some basic knowledge, and when this basic knowledge is not too hard for me to think about, I can then have the extra mental energy to understand "beauty."
Today is the last day. Recently, I've had some pressure from life and work, which made me a bit tired of learning. My current state is probably not doing fifty [XP] daily anymore, but three hundred and fifty within a week. Overall, I think this is not good. After all, based on what I know about how to learn scientifically, I certainly know that it's better to learn less but spread out, rather than learning a lot at once. But there's not much I can do about it; continuing to learn is the most important thing no matter what. When I really don't have time on weekdays, I can only move it to study hard on Saturdays and Sundays, which sort of meets the requirement of three hundred and fifty per week.
Looking back over the month, I found I've really learned a lot of things: unit circle, trigonometric functions, basic geometric triangle knowledge, logarithms, factoring, complex square roots, inverse functions, complex numbers, solving inequalities. These are indeed very basic things, probably not harder than junior high school level, but compared to the little math knowledge I had before, it's already a big step forward. I had read many traditional textbooks before, or followed online courses, but never for more than a week. I really didn't expect to be able to learn so smoothly and continuously for a month while still meeting the weekly XP task – not only without strong feelings against it (if you don't count tiredness due to stress, which I think has nothing to do with MA), but I even plan to continue paying to start the second month. You could say that the scientific results of MA's designers have worked well for me.
Traditional things don't work for me, or even have a negative effect. I thought about finding a tutor, but unfortunately, I couldn't afford it, and it even once made me feel like maybe I would never be able to learn math in my life. And now I've finally found the "only" method I can learn with. Forget just the second month, I think I'll continue for a year, as I don't have many other options. I used to think the path of mathematics was not something ordinary people like me could climb, but now I suddenly have the ambition to "work through" MA and learn everything I can. That's right, I've only learned the basics; maybe someday I'll hit a limit with abstract thinking. But undoubtedly, at my current level, I'm probably still far from my limit. I am very grateful to the designers of MA. If it weren't for them, my confidence in mathematics would likely be very difficult to rebuild. To learn for a long time, motivation and confidence are absolutely essential. Here, MA not only gave me knowledge but also gave me the hardest thing to buy – confidence in learning.
[1]
[2]
[3]
