# GridBot — 5-Day Curriculum: *from zero to a trained soccer-player bot* A week-long (or five-session) plan that takes a kid with **no coding experience** to a robot they **trained themselves** to play soccer — and, just as importantly, an understanding of *when to write the rules and when to let the robot learn them*. - **Audience:** ages ~7–13, one kid per **$10 CYD** (or pairs sharing one). **Got a younger kid (~6–8)?** Start with the gentler **[Younger Kids Path](YOUNGER-KIDS-PATH.md)** — the same five days in 15-minute chunks, one big idea a day, with the exact sentences to say (a prequel you can graduate from, or a parallel calmer route). - **Time:** **5 days, 1–2 hours each** (≈ 7–10 hours total). - **Setup:** none — it's all on the device. No internet, no accounts. For two-device days you just need two CYDs in the same room. - **This is the one-page plan.** Want everything in a single self-contained guide you can teach straight from — every program, question, and checkpoint inline, no flipping between docs? Use **[COURSE.md — The Complete Course](COURSE.md)**. The other companion docs go deeper on specifics: the [README](../README.md) (the what), the [HANDBOOK](HANDBOOK.md) (the full teacher guide, a page per lesson), and the [Hand-Coding Guide](HAND-CODING-GUIDE.md) (writing bots by hand). > **How features unlock.** Powers arrive as you climb campaign levels — **jump (Lv 6) · loops (Lv 10) > · sensing `if`/`until` + the Arena (Lv 15) · functions (Lv 20) · NeuroBot training (Lv 28)**. So the > campaign *is* the spine: each day plays a few levels to earn the day's new power, then uses it. Pace > is ~5–6 levels per session; a kid who's flying can race ahead, a kid who's stuck can slow down — > the **2-/3-star** scoring rewards coming back to a level with a cleaner program. > > **The game teaches the moment too:** the first time a kid reaches the level that introduces a power > (Move at Lv 1, then Jump/Repeat/Sense/Functions), its CodeLab lesson **opens automatically, once** — > so even without an adult driving, no power arrives unexplained. Doing a lesson from **Learn → CodeLab** > first counts (it won't replay), and every level's intro card keeps a **"Learn it >"** button to re-watch. --- ## The week at a glance | Day | Theme | New power | They can… | The big idea | |---|---|---|---|---| | **1** | 🚗 **Drive** | move / turn / **jump** | steer the robot through a maze by hand | a **program is a sequence** | | **2** | 🧠 **Think** | **loops** + **sensing** (`if`/`until`) | write **one** program that solves mazes it's never seen | a **rule generalises** (an *algorithm*) | | **3** | ⚔️ **Compete** | **functions** + the **Arena** | hand-code a **battle** bot and a **soccer** bot | clear rules win — *until they don't* | | **4** | 🤖 **Learn** | **NeuroBot** (Teach / Q-Learn / Evolve) | **train** a brain instead of writing rules | when a job is feel, not rules, **let it practise** | | **5** | 🏆 **Champion** | the trainer's knobs + the **Room** | train a soccer **striker** and run a **tournament** | iterate, measure, and *match the method* | ### Same week, for ages ~6–8 If you're running this with a younger kid, here's the **one-big-idea-a-day** version — the full **[Younger Kids Path](YOUNGER-KIDS-PATH.md)** has the 15-minute quest cards, grown-up scripts, and a Parent Rescue Guide. The right column is what to **skip until later** at that age. | Day | Big idea (say this) | Younger Kid Version | Skip Until Later | |---|---|---|---| | **1** | "The robot follows my list." | Make a list of moves, fix one wrong turn | CSTA terms, star chasing | | **2** | "The robot can check before it moves." | Check wall, then turn | the word "algorithm", *why* it generalises | | **3** | "Rules need an order." | First matching rule wins | functions, win-rate tables | | **4** | "Training means trying and getting better." | Press **Teach**, watch it improve | Teach/Q-Learn/Evolve taxonomy | | **5** | "We test which bot is best." | Try bots, keep the one that wins | statistics, "one match can fool us" | --- ## Day 1 — 🚗 Drive: *make it move* **Goal:** the robot obeys a list of steps you write, and you solve a maze by hand. 1. **Pick a player** (avatar + name) — this saves their progress. 2. **CodeLab → "1 · Move"** (Learn → CodeLab): tap **Run**, watch forward/turn/forward reach the goal. 3. **Play campaign levels 1–9.** Drag command blocks, tap **Run**, read the failure, fix one block, re-run. **Jump unlocks at Lv 6** — use it to clear a pit. 4. **Star chase:** beat a level, then try to beat it again with **fewer blocks** for 3 stars. **Ask:** *"What happens if two steps swap places?"* (forward-then-turn ≠ turn-then-forward — **order matters**). **Checkpoint:** the kid can plan a turn-by-turn path and debug a wrong turn. --- ## Day 2 — 🧠 Think: *make it solve mazes it's never seen* **Goal:** the leap from *memorising steps* to *a rule that adapts* — the heart of the whole course. 1. **CodeLab → "3 · Repeat"** (loops unlock at **Lv 10**): replace five forwards with one loop. *Why is a loop better than copy-paste?* (efficiency **and** readability.) 2. **CodeLab → "4 · Sense"** (`if` / `until` unlock at **Lv 15**): the robot **reacts** to what it senses. Build the **wall-follower**: `until goal { if wall { turn left } forward }`. 3. **The "aha":** run the *same* wall-follower on **three different mazes** without changing a block. It solves all three. *Why?* Because it's a **rule**, not a memorised path. 4. **CodeLab → "5 · Functions"** (Lv 20): name a group of steps and reuse it — the program gets shorter *and* clearer. **Ask:** *"This program, unchanged — could it solve a maze you've never opened? Why?"* **Checkpoint:** the kid explains *why* the wall-follower generalises. *(This is the curriculum's hinge — don't rush it.)* --- ## Day 3 — ⚔️ Compete: *hand-code a fighter and a striker* **Goal:** use sensing to build **Arena** bots by hand — and feel where hand-coding starts to strain. Work from the **[Hand-Coding Guide](HAND-CODING-GUIDE.md)** (it has on-device screenshots of each program). 1. **Battle (Sumo):** build the **hunter** — a priority list: *zap if the foe's ahead; never charge a pit; turn to face the foe; otherwise close in.* Fight **vs Computer** and win. **Reorder the rules** (put `forward` first) and watch it get dumb — *order is the whole lesson.* 2. **Soccer:** build the **dribbler** — `if ball^` (touching the ball) check the **net** and circle behind it before pushing, so a shove sends the ball toward the *opponent's* goal. (You can't score on yourself — shove it the wrong way and your robot auto-turns around with it — but you still have to get behind it to score.) You can also flip the ball on demand with the **`zap`-swap**: `if ball^ { zap }` makes your robot **trade places with the ball** to turn it around. 3. **The honest moment:** the hand-coded striker *plays*, but a good opponent beats it. Hand-coding nailed the maze and the fight; soccer needs **finesse** that's hard to write down. **Ask:** *"Why is 'follow the wall' easy to write as a rule, but 'dribble past a defender and pick your corner' is not?"* **Checkpoint:** a working hand-coded soccer bot **and** the realisation that it's not enough — which *earns* Day 4. --- ## Day 4 — 🤖 Learn: *stop writing the rules, start training them* **Goal:** reach **NeuroBot** (unlocks at **Lv 28**) and train a brain from examples and reward. 1. **NeuroLab lessons** (Learn → NeuroLab) — small enough to *watch*: **One neuron** (it guesses, sees the error, nudges), **Perception** (the 10 senses the brain gets), **Many actions** (go/turn/jump), **One brain, many jobs** (the same brain taught **maze → battle → soccer** in a row — only the senses and the teacher change), and **The Right Tool** (*match the method to the problem*). 2. **Arena → Train a fighter → Soccer.** Try all three engines and watch the **learning-curve**: - **Teach** — *imitate an expert dribbler* (sharp in seconds). - **Q-Learn** — *reward* a goal; no teacher, just "did it score?". - **Evolve** — *the best brains breed and mutate.* 3. **Code vs brain:** field a **trained** soccer brain against yesterday's **hand-coded** one. **Ask:** *"It learned mazes, then fighting, now soccer — what stays the same about its brain, and what changes?"* (the network is identical — `10 → 8 → 5` — only what it senses as the goal and how it's rewarded changes.) **Checkpoint:** a trained brain that out-plays the hand-coded bot at soccer. --- ## Day 5 — 🏆 Champion: *train a striker, then a tournament* **Goal:** refine a real soccer **striker**, then crown a champion across the class. 1. **Train a striker properly:** **Teach** a dribbler first — it's strong in seconds. *Keep that Teach base* and only adopt a refinement if a **rematch proves it's better**: in our measured bake-off, Teach (20%) was the strong base, refining **against the real opponent** (Teach→Evolve, 47%) helped most, while **Q-Learn *hurt*** (0%/4%, below Teach). (No own-goals to train away: shove the ball toward your own net and the game auto-turns your robot around with it, so training is purely about *scoring on the opponent*.) **Save** your striker. *(Tune the **Knobs** — learning rate, rounds, explore — and watch the loss fall, or thrash.)* 2. **Level-up loop:** lost a match? **Save foes** (copy the winner into your library), train against *it*, and **rematch**. Because the Arena is **deterministic**, a rematch is exact — so improvements are real, not luck. 3. **Tournament** — the capstone: - One device: **vs Computer → Tournament** (a Cup bracket or Ladder of your saved fighters). - The class: **Arena → Radio → Room** gathers *every* nearby CYD's striker into **one shared bracket** — every screen replays the identical Cup from a shared seed and crowns the **same champion**. No central scoreboard needed. **Ask:** *"Your hand-coded bot vs your trained bot — which would you bring to the tournament, and why?"* **Checkpoint (the finish line):** a **soccer-player bot the kid trained themselves**, fielded in a real bracket — and a kid who can say *when* they'd write the rules and *when* they'd train them. --- ## What they walk away with - **Coding:** sequence → loops → conditionals → functions → debugging (CSTA-aligned; see the HANDBOOK §9 standards map). - **AI literacy (AI4K12 "Five Big Ideas"):** what a network *senses*, how it *learns* (imitation, reward, evolution), and the meta-skill of **matching the method to the problem** — the difference between a *rule* and a *trained* behaviour, felt in their own hands. - **A robot they made and trained** that plays soccer — and the confidence that AI isn't magic. ## Teacher notes - **Pairs work great:** one drives, one reads the failure aloud — narrating the *read → guess → change one thing → test* loop is the real computational-thinking lesson. - **Falling behind on levels?** The lessons (CodeLab/NeuroLab) teach the *concepts* independently of the campaign grind — run the lesson even if a kid hasn't unlocked that level yet, then let the campaign reinforce it. - **Going long?** Day 5 alone can fill a "science fair" afternoon: everyone trains a striker, then a Room tournament decides it. - **Every number a kid will see** (this bot beats that one) is **reproducible** — the Arena is deterministic, so results are comparisons, not coin-flips.