TAI-VEX Override HS: Design Award, Tournament Finalist, Skills 2nd Place

On October 25th, we went to our team's first tournament after the MOA Signature Event. We really improved a lot since the MOA Signature Event, and we were able to perform consistently in the tournament.

Although we did not perform well during the MOA Signature Event, we learned a lot from the top teams there. After MOA, I worked really hard to combine all of the best designs from different top teams in MOA into our Third Generation Robot.

After finishing the CAD, the robot was built pretty quickly. We were able to finish the robot in early September, giving us a whole month to practice driving and code the autonomous routes.

In both September and October, our robotics lab held scrimmages to help us familiarize ourselves with playing tournament matches. Not only did we win both scrimmages, but we also improved our strategy, communication, and driving skills through the matches.

I also spent quite a lot of time tuning the autonomous program and practicing driver skills to improve our skills score (which was 38 points in MOA).

Going to the competition, we felt much more confident and prepared compared to MOA. 

Since we could stack quite efficiently, we were pretty good at tournament matches. We won all but one qualification match and ranked 4th on the qualification ranking. 

Luckily, our sister team 66799T was ranked 1st and they selected us as their alliance partner. So, we went into the elimination rounds as the first seed.

In the Semifinals, we faced our sister team 66994D. I made the wrong decision of going to steal opponents' toggles instead of protecting ours, letting opponents steal our toggle. We would've lost if the opponents weren't disqualified for lifting our teammate.

We were able to win Final 1 with a really big margin after stealing opponents' toggle.

In Final 2, the opponents switched to a more conservative strategy. In the last 5 seconds, our opponents activated their toggle extension mechanism and touched our toggle, making it worth 0 points and costing us the match.

Final 3 was the closest match. In the beginning of the match, I accidentally hit a neutral goal, knocking off 2 pins. This time, I successfully defended our home toggle, but we left all other toggles undefended. Although we still won by 15 points, we were disqualified due to the earlier neutral goal descoring.

In the skills judging, we aimed for 128 points in driver skills and 108 in auto skills. I ran slowly in the first driver attempt and got 118 points, a solid start. Then we failed the first auto attempt because I ran the test route instead of the full one. The second and third auto attempts both failed because we weren't able to flip the second toggle. Still, we got 50 points in auto, much better than the 10 points we got in MOA.

Speaking of judging, we did pretty well since we had a pretty complete notebook and a well-organized team interview. Luckily, our work was recognized by the judges, winning us the Design Award.

We were really close to winning the tournament, so we were all quite depressed after the finals. But the tournament that can actually win us a worlds spot is the PAS-VEX Signature Event in December. So we have no time to feel discouraged, and we must quickly build our next generation robot to overcome the challenges we have observed throughout this competition.

Overall, it was a fun and rewarding experience. The hard work after MOA really did pay off. Along with the experience gained during this tournament, we shall do even better in upcoming competitions.

How I Lost at the MOA Signature Event (And Why I Don't Regret It)

To start my team's journey for this VEX competition season, we participated in the University of North Dakota Signature Event at Mall of America. It was definitely a unique and fun experience that I would like to talk about today.

After deciding to compete back in May, we started brainstorming, designing, and building robots. The robot we used in Mall of America was our second generation robot.

Our first generation robot was quite simple, it focused on quick scoring and short cycle time. However, it wasn't that consistent and it didn't have the ability to take objects from the ground, which we thought was quite important.

So, we designed a second generation robot. The design was really unique and bold, we were the first team in the world to build something like this. But this also meant we were the first team to face the crazy amount of new problems on the robot structure.

For example, the main problem we faced was the cup hitting our bottom lift bracing. Theoretically, if the claw could grab onto the cup perfectly upright, they wouldn't collide. But in reality, even the slightest offset causes the jam. Although we developed an anti-jam process later, it drastically damaged our scoring efficiency.

Going to the competition, we knew our robot wasn't as efficient as the others. So the only way we could win was through clever strategies and smooth robot driving.

In the 9 qualification matches, we won 5. Unfortunately, one of the win was given to the opponents because we accidentally knocked down a single pin and received a major violation. 

We were also quite unlucky in the losing matches. In one match, we were pinned on a goal for 39 seconds without the opponents getting a violation. In another, our teammate had their battery unplugged in the middle of the match.

If we were lucky and won the two matches we could've won, and didn't get disqualified on the winning match, we would've been at ranking 20. Instead, we were ranked at 84 at the end of qualification matches.

Despite our best efforts of convincing teams to select us for elimination matches, we were ultimately left unpicked. Frankly, we were not qualified enough to reach the elimination rounds anyway.

Luckily, we didn't end up with nothing for this trip. The biggest treasure we have gained was the increase in our team's visibility. Through the robot reveal and posts on Discord servers, we really got quite a lot of people to know us for the first time.

Also, we were honored to be interviewed by FUN Robotics Network. We were able to briefly introduce our unique lift build and our robot simulator that helped us designed our complicated robot code. 

The idea of a robot simulator was born during the tedious robot building and tuning period. I wanted to start practicing driving and writing the robot program while the robot build was in progress. So I built the simulator that could simulate the robot movement on the field, and more importantly, the lift and arm movements. As a result, we had a functioning pose machine and control algorithm for our lift and arm only a few days after the finish of our robot build.

So, although it is true that we might've performed better with our first generation robot, I don't regret the decision of competing with the second generation robot. Without the unique design, we wouldn't have attracted nearly as much attention as we did. Without the difficult robot tuning process, we wouldn't have learned how to build a continuous lift well. 

With this experience, I'm sure that we could build even better robots and win in future tournaments. So, this event is only the beginning. We shall continue working towards Worlds qualification.

 

JLPT N1 通過

最近我考 N1 的成績出來了，而我今天我想分享日文檢定 N1 的準備過程和一些心得。

考試時間是七月初，而我是從五月底開始準備的。其實我因為一開始覺得 N1 非常困難，所以想要逃避，但後來爸爸鼓勵我努力試一次，畢竟我六月空閒時間比較多，如果幸運考過的話以後就不需要煩惱了。

根據姊姊之前考 N1 的經驗，寫考古題是最有效的方法，所以我也是用這個方法準備。

我第一次寫考古題，文字語彙和讀解大概都只對一半，所以我去問了 N1 讀解滿分的姊姊，我覺得最大的改變是心態，仔細讀過的話會發現題目沒有這麼難，只要能找到關鍵的句子就能夠得到答案。經過大概五回考古題，我在讀解部分進步到可以對八成左右。

至於文字語彙的部分，我覺得真的比較難短時間內練起來，只能透夠持續寫題目來學更多單字和文法。聽解的話也蠻困難的，但我覺得能夠在紙上很快記下重點幫助很大，尤其是後面重點很多的題目。

考試當天，我記得文字語彙很難，幾乎都是我沒遇過的單字，所以有點害怕沒拿到 19 分而不通過，還好最後運氣不錯，有拿到 26 分。讀解的部分倒是蠻簡單的，我大部分的分數都是在這裡拿到，最後得到 53 分。聽解我記得很難，大概只有三分之一的題目是我確定的，最後得了 33 分。

所以三個項目加起來我得了 112 分，只比通過標準 100 分多一點。如果我少猜對幾題可能就要年底重考了。

總結來說，寫考古題幫助很大，但我實際考試運氣也不錯，才能驚險通過。我想要感謝提供經驗跟建議的姊姊，幫助真的很大，還有父母的堅持讓我有動力在知道考試很難的狀況下努力準備。

2026 GJMAT Gold Award

Recently, I took GJMAT, an India-based mathematical competition. It was an online competition and overall pretty interesting.

In total, there were 25 problems to solve. The first 10 were only worth 2 points each, problems 11~20 were worth 4 points each, and the last 5 problems were worth 8 points each.

When the exam started, I started from the first problem. However, I got stuck when I hit the third problem. So I decided to jump to the last problem and work my way backwards to earn the most points.

Surprisingly, I was able to complete most problems. I still missed a few problems, but I had confidence on around 20 of the problems. 

Not only was the math itself hard, but understanding the Indian English was also a challenge. Sometimes I couldn't fully understand the problem quickly. 

Recently, the results came out, and I was lucky enough to receive the gold award. Unfortunately, they didn't provide my scores and ranking.

Overall, the GJMAT was a fun experience. I had fun working on the problems, and ultimately I won the gold award. Sadly, I think GJMAT is only available up to grade 10. So this was one of the last years I could join GJMAT.

My VEX World Championship 2026 Experience

This blog documents my trip to the VEX World Championship in St. Louis. I was able to experience a lot, meet the true top teams in the world, and learn from them.

This was my first year competing in the VEX V5 Robotics Competition, and along with my great teammates, we were able to qualify for the World Championship in the PAS-VEX Signature Event in December.

The VEX World Championship was featured over 800 teams from around the world, representing a scale entirely different from any other competition we have attended.

Pre-Competition Practice

Before the competition, we held practice matches in a conference room we rented and invited teams to join and practice together. We were honored to have both 334C and 334U join the games. They were the best teams in the world, on a completely different level compared to all teams in Taiwan.

Both 334C and 334U had two wings, and they were really good at using them. So we couldn't win in a wing battle against them at a long goal. That's why we switched strategy and avoided directly fighting them. We would score mainly in the center goals, and only dump on the long goals at the end. As a result, we were able to win a match against them by a slim margin. 

Being able to practice with these top teams really changed me. I realized that they were so much better than us. In the next season, we shouldn't aim only for the champions in Taiwan, but also aim to become a world top team.

Competition + Getting Sick

Unfortunately, I became really sick on the first day of the competition. Therefore, I only competed in the afternoon. We won the practice match but lost our first qualification match.

I returned early on the first day and rested in the morning of day 2 until I was called to the convention center because we had a team interview. So despite being pretty sick and having almost no voice, I raced to the convention center and completed the team interview along with my teammates.

For day 2, we won 3 matches and lost 2 in total. Our ranking was around the 30s, which was pretty bad. At night, we fixed the robot to make our intake less likely to overheat. 

We had a really easy match schedule on day 3, and along with our fixed robot, we were able to win all 5 matches. We ended up 15th on the qualification ranking. This guaranteed us a seat in the elimination, but we still had to talk to top ranking teams to seek alliance.

Eliminations

Finally, after 8 alliances, 18961B picked us and we accepted. Forming the 9th seed. 18961B was a team from China and they had a pretty nice robot. So we practiced auto together and headed to the elimination rounds.

The first match was the round of 16; we faced 13722V and 3632A. We fought a pretty tough match, but we ultimately won with a 15-point margin.

The quarterfinal round was really challenging because our opponents were the first seed in our division, 11101X and 10222R. Luckily, we were able to take 2 goals and our alliance partner got a park. We ended up winning the quarterfinal by 12 points.

The semifinal round was even harder, with 81988E and 15442A as our opponents. 81988E had the ability to block both center goals, making it hard for us to win the center. During the match, they blocked the center goal. Unfortunately, our teammate got stuck with them in the center goal, and we lost the match by a good margin.

Conclusions

Although we were stopped in the semifinal, we were able to get the Create Award in our division. We were really glad that we didn't have to go home empty-handed. 

Overall, I really enjoyed this year's VEX World Championship. Practicing with 334C and 334U really made me realize how good other teams were. We also did quite well in the tournament as we won 9 out of 12 qualification matches. Forming the 9th seed, we defeated the 8th seed and 1st seed and went into the semifinal round. Finally, the Create Award recognizes our work on the notebook, interview, and robot.

I had a great experience. Unfortunately, I was really sick for most of the competition. I'm sure I would've had a lot more fun if I was 100% healthy. 

Finally, I have to thank my teammates who made the team so well-rounded. Thanks to our sister teams for constantly practicing with us. Thanks to our coach who taught us a lot throughout the season. Thanks to all of the teams we met during the competition for making VEX Worlds fun. Thanks to the parents who were willing to support us no matter what.

2026 AIME II 8/15

This is my first year competing in AIME. I thought it would be really hard, but I ended up with 8 questions answered correctly, way above my expectations before taking the test.

After getting an 118.5 on the AMC 10A last year, I qualified for this year's AIME. To prepare for the AIME, I tried looking at some of the problems from 2025 and 2024. It turned out that I could barely solve any of the problem. That's why I thought I was cooked for this year's AIME.

For the next few weeks, I decided to start taking one test every week, starting from 1983, so I could build up confidence and actually get some questions correct.

On the day of the test, my expectation was that I could get 3 to 4 questions right. But when I really started working on the problems, they turned out being really easy, at least a lot easier than the previous problems.

The first 5 problems were quite easy and I completed them quickly. Then I continued finishing questions 6 to 8. After question 8, the problems became so difficult that I couldn't even think of a way to approach them. Just when I was desperate, I saw question 15, and it was surprisingly easy to solve. 

The whole test was 3 hours long, and I only had 15 problems to work with. So after finishing all problems that I could solve, I started drawing and writing random things, including names of small Japanese political parties. For the last 30 minutes, I pretty much slept until the bell rang.

Fast forward to a few weeks ago, the scores came out. I got 8 correct out of the 15 problems. It turned out that I missed question 6 unfortunately. Finally, out of the 270 contestants in Taiwan, I was in the top 41. There was one person who scored a 15, which is totally unbelievable.

Overall, it was a really fun experience to compete in the AIME. I hope I could still qualify for the AIME and do even better next year. 

自學第八年 回顧與自評 / Yu

回顧期間：2025 / 4 - 2026 / 3   填寫日期：2026 / 3 / 20

1. 這是自學的第八年，對鎮宇來說，已經要開始考慮申請大學了。想先請鎮宇介紹一下，過去一年，你在學科方面的努力，包括：台灣國民教育教材學習了哪些，以及開始讀的 AP 課程等。

過去一年，我把國民教育的課程裡面想學的都學完了，包括以下科目：

• 高中數學二上、二下

• 高中選修物理 I

• 高中地理二上

• 高中歷史二上

• 高中公民二上

另外，我也開始準備 AP 的各科，今年我打算考六個科目，包括：

• AP Calculus AB

• AP Statistics

• AP Physics 1

• AP Human Geography

• AP World History: Modern

• AP Japanese Language & Culture

我覺得國民教育的內容先學完很好，讓我學習 AP 科目變得輕鬆很多，而 AP 各科我現在讀完內容，開始大量考模擬考了，目前我在模擬考都可以拿到 5 分。

2. 今年你也投入了 VEX V5 的比賽，從 IQ 轉 V5，心得很多。請跟我們分享一下，這個賽季目前的成績，以及這一年來你自己感受到的，有哪些很重要的進步跟改變？

今年是我第一年比 VEX V5，跟之前比的 VEX IQ 非常不一樣。VEX IQ 因為是塑膠零件，很容易拆裝，基本上不用思考過就可以開始組；但 VEX V5 主要用鋁，每個接點都要鎖螺絲，組機器消耗的時間多很多，所以通常要先用電腦把機器畫出來。

團隊分工的方面也改變很多，VEX IQ 基本上我一個人就可以做完全部的事情，但 VEX V5 有更多方面，所以一定要分工合作。

這個賽季到現在，我們參加了三場主要的比賽，也拿到很多獎項：

• TAI-VEX HS：Excellence Award、Tournament Champion、Skills Champion

• PAS-VEX Signature Event：Tournament Champion

• VEX Taiwan Open HS：Create Award、Tournament Finalist

3. 你最在行的數學，這一年來你參加了不少的競賽，也包括許多國際賽，請介紹一下過去一年所獲得的獎項，並選一個你覺得最有意思的，分享一下你的心得體會。

以下是我這一年來參加的數學比賽和成績：

袋鼠數學：滿分、世界第 1 名

GJMAT：世界第 12 名

GJMOC：世界第 7 名

AMC 10：傑出、Certificate of Distinction（前 5 %）

AIME：8 分

其中最重要的應該是 AMC 10 和 AIME，其中 AIME 是 AMC 10 排名前 2.5% 才能考的，目前成績報告書還沒出來，所以不確定 8 分是什麼樣的成績，但我已經很滿意了。

 

4. 語言部分，今年你通過了 N2考試，請分享一下你的準備經驗跟結果。未來一年，會再繼續挑戰 N1 嗎？

我寫了一本 500 題單字的書，並反覆複習，增加了很多單字量，但是考試中好像只有出現一兩次。所以感覺寫模擬考題比較有效，至少知道自己的分數和考試的節奏。

我已經報名了今年 7 月的 N1 考試，就先考考看，沒有通過也可以，之後有空的時候再來用模擬考衝刺。

5. 過去一年，以大型語言模型 LLM 技術達成的 AI，改變了世界，越來越普及，也越來越聰明。請分享一下，在過去一年，你最常用那個 AI 工具，以及應用在哪些方面。你認為，AI 對你做為學生的學習需求上，是有用的嗎？

我主要使用的是 Google 的 Gemini AI，感覺他在各個方面都很強。我在各個方面都會問他，除了 AP 學習上遇到的問題之外，機器人程式遇到的困難、去日本旅遊看到好奇的事都可以問他並得到完整的答案。

以往遇到困難只能想辦法 Google 搜尋，但現在問 AI 比較快也比較準。當然，如果問資訊量比較少的領域，AI 很容易給錯誤資訊，所以重新驗證資訊的準確性也非常

6. 除了學科、機器人、數學競賽之外，你也自行吸收了很多網路上的知識，尤其 YouTube 的各種英文頻道。請選三個你覺得過去這一年，讓你覺得很有意思的影片或者頻道，並分享你的體會。知識、遊戲、攻略、迷因、流行文化、政治歷史等等的都可以喔！

一整年來，我看的影片有很多種類，演算法推薦的影片也會隨時間改變，但是以下是我最有印象的幾個頻道：

Jet Lag: The Game：這個頻道會去世界各地玩遊戲，以整個國家作為地圖，然後搭火車或飛機到達各個地方，我覺得很有趣。順帶一提，他們最新的一季會在臺灣玩。

Luke does robotics：他是一個美國大學生，他會做關於 VEX V5 比賽的影片，我去年就是看他的影片才開始對比賽的策略感興趣的。

Denys Davydov：他是烏克蘭人，以前是機師，現在他每天更新烏克蘭戰況，讓大家比較知道前線的戰況跟局勢的發展。