The Completionist's Handbook — How to Collect All 18 Species and Document Your Buddy Zoo
[01]The Collector's Itch
You've hatched your first buddy. Maybe it's a Common Duck, maybe a Rare Ghost. Either way, you've seen the species catalog — 18 creatures, each with its own ASCII sprite, backstory, and personality. And now a question gnaws at you: what would I get with a different UUID?
That question is the collector's itch, and it has a formal name in mathematics: the Coupon Collector's Problem. First studied by Euler and formalized by Flajolet in 1992, it asks a deceptively simple question: if you draw randomly from n equally likely items, how many draws until you've seen every item at least once?
For Claude Buddy, n = 18. The species roll is a uniform pick from the SPECIES array — every species has exactly 1/18 ≈ 5.56% probability, regardless of rarity. This means the math is clean, the simulations are precise, and the strategy is real. Let's build your zoo.
[02]The Math: Coupon Collector's Problem
The expected number of trials to collect all n equally likely coupons is:
E(T) = n × H(n)
where H(n) is the n-th harmonic number: H(n) = 1 + 1/2 + 1/3 + … + 1/n.
For 18 species: H(18) = 3.4951, so E(T) = 18 × 3.4951 ≈ 63 trials. The standard deviation is approximately 23, meaning most collectors will finish between 40 and 86 trials.
The intuition is beautiful: your first species is guaranteed (1 trial). The second requires 18/17 ≈ 1.06 trials on average. But the last species — when you already have 17 — requires 18/1 = 18 trials on average. That final holdout is where patience is tested.
| Species # | Already Collected | Probability of New | Expected Trials | Cumulative Expected |
|---|---|---|---|---|
| 1st | 0 | 18/18 = 100% | 1.0 | 1.0 |
| 2nd | 1 | 17/18 = 94.4% | 1.06 | 2.06 |
| 5th | 4 | 14/18 = 77.8% | 1.29 | 5.56 |
| 10th | 9 | 9/18 = 50.0% | 2.00 | 14.69 |
| 15th | 14 | 4/18 = 22.2% | 4.50 | 33.19 |
| 17th | 16 | 2/18 = 11.1% | 9.00 | 45.94 |
| 18th | 17 | 1/18 = 5.56% | 18.00 | 63.94 |
Notice how the last three species alone account for 30 of the 64 expected trials. This is the collector's curse: the closer you get, the slower it feels.
[03]100,000 Simulations: The Real Numbers
Theory is elegant, but we ran 100,000 Monte Carlo simulations to get the exact probability distribution. Here are the milestone numbers every collector needs:
| Milestone | Trials Needed | Meaning |
|---|---|---|
| Minimum observed | 21 | Luckiest possible (top 0.01%) |
| 25th percentile | 49 | 1 in 4 collectors finish by here |
| Median (50%) | 59 | Half of all collectors finish by here |
| Mean | 63 | Mathematical expected value |
| 75th percentile | 73 | 3 in 4 collectors finish by here |
| 90th percentile | 91 | Only 10% need more than this |
| 95th percentile | 103 | Unlucky but not impossible |
| 99th percentile | 131 | Extremely unlucky |
| Maximum observed | 299 | Worst case in 100k sims |
Key takeaway: if you've tried 100 UUIDs and still haven't found all 18, you're in the unluckiest 5%. But you're not broken — the algorithm is working exactly as expected. Keep going.
[04]The Five Collector Tiers
Not every collector has the same goal. We've defined five tiers of increasing ambition, each with its own math and strategy:
| Tier | Goal | Target Count | Expected Trials | 90% Confidence |
|---|---|---|---|---|
| 🌱 Explorer | See 9 of 18 species (50%) | 9 | ~22 | ~30 |
| 📦 Collector | All 18 species (any rarity) | 18 | ~63 | ~91 |
| ⭐ Rarity Hunter | All 5 rarity tiers | 5 | ~106 | ~231 |
| 💎 Combo Seeker | All 90 species × rarity combos | 90 | ~9,000+ | ~15,000+ |
| ✨ Shiny Completionist | At least one shiny of each species | 18 shinies | ~115,000+ | ~180,000+ |
The jump between tiers is exponential. Going from Collector to Rarity Hunter roughly doubles your effort. Going from Rarity Hunter to Combo Seeker multiplies it by 40×. And Shiny Completionist? That's a lifetime project — each shiny-species combo has only a 1/1,800 chance per trial.
[05]Strategy 1: The Systematic UUID Sweep
The most efficient collection strategy is a systematic sweep using the Buddy Checker at claudebuddy.art. Here's the method:
- Start with any string. Type "test-001", "test-002", "test-003" into the checker. Each string produces a deterministic buddy.
- Log every result. Create a spreadsheet with columns: Input String, Species, Rarity, Shiny, Hat, Eye. The checker shows all of these instantly.
- Track your gaps. After 30 trials, you'll likely have 13-14 species. Identify the missing ones.
- Keep going until 18/18. The last 2-3 species will take the most patience. Don't change your naming scheme — just increment.
Why the Buddy Checker? Because it uses the exact same algorithm as Claude Code: FNV-1a hash → Mulberry32 PRNG → sequential rolls. Every result you see on the website is mathematically identical to what /buddy would produce. You can check hundreds of UUIDs in minutes without touching Claude Code.
[06]Strategy 2: The Brute-Force Script
For the programmatically inclined, you can write a script that runs the buddy engine against thousands of inputs. The algorithm is public and deterministic:
// Pseudocode for species collection sweep
const collected = new Set();
let trials = 0;
while (collected.size < 18) {
const uuid = `collector-sweep-${trials}`;
const buddy = rollBuddy(uuid);
collected.add(buddy.species);
trials++;
if (!seen[buddy.species]) {
console.log(`NEW: ${buddy.species} found at trial ${trials}`);
}
}
console.log(`All 18 collected in ${trials} trials`);
In our tests, this approach consistently finds all 18 species in 50-80 trials, matching the theoretical prediction. The script can also log rarity, shiny status, and hat for each find, building your complete zoo database automatically.
Pro tip: Use a prefix that means something to you. "alice-001", "bob-001", or even dates like "2026-04-16-001". Each input is a permanent, reproducible identity — you can always re-check it later.
[07]The Rarity Bottleneck
If your goal extends beyond species to rarity tiers, the math changes dramatically. Species are uniformly distributed (1/18 each), but rarities are heavily skewed:
| Rarity | Weight | Probability | Expected Trials for First |
|---|---|---|---|
| Common | 60 | 60% | 1.67 |
| Uncommon | 25 | 25% | 4.00 |
| Rare | 10 | 10% | 10.00 |
| Epic | 4 | 4% | 25.00 |
| Legendary | 1 | 1% | 100.00 |
Our 100,000 simulations for collecting all 5 rarities show:
- Median: 75 trials (half finish by here)
- Mean: 106 trials (skewed by legendary bottleneck)
- 90th percentile: 231 trials
- 99th percentile: 463 trials
The bottleneck is always Legendary. You'll likely collect Common, Uncommon, Rare, and Epic within your first 30 trials. Then you'll spend 70+ more trials waiting for that 1% Legendary to appear. This is normal — the math demands it.
[08]Building Your Buddy Zoo
A true completionist doesn't just collect — they document. Here's how to build your personal Buddy Zoo:
The Zoo Spreadsheet: Create a grid with 18 rows (species) and 5 columns (rarities). Each cell records the UUID that produced that combination. Mark shiny variants with a star. This becomes your permanent collection record.
| Species | Common | Uncommon | Rare | Epic | Legendary |
|---|---|---|---|---|---|
| Duck | test-003 ✓ | test-047 ✓ | — | — | — |
| Cat | test-011 ✓ | — | test-089 ✓ | — | — |
| Dragon | — | test-022 ✓ | — | test-156 ✓ | — |
| … | … | … | … | … | … |
The Share Card: Use the Buddy Checker's built-in share card feature to generate a 1200×630 image of your best finds. Post them to the community with the UUID so others can verify.
The Verification Principle: Every entry in your zoo is independently verifiable. Anyone can type the same UUID into the Buddy Checker and confirm your claim. This is the beauty of deterministic systems — there's no faking a Legendary Shiny Dragon.
[09]The Last Species Problem
Every collector hits the same wall: 17 out of 18 species found, and the last one refuses to appear. This is mathematically inevitable — the expected wait for that final species is 18 trials, but the variance is high. You might get lucky in 5 tries, or you might need 50.
Here's the probability of finding your last species within k additional trials:
| Additional Trials | Probability of Finding It |
|---|---|
| 5 | 24.5% |
| 10 | 43.0% |
| 18 | 63.2% |
| 30 | 81.3% |
| 50 | 93.8% |
| 80 | 98.9% |
The formula is simple: P(find within k trials) = 1 − (17/18)^k. Each trial has a 5.56% chance, and the attempts are independent. If you've been stuck for 30 tries, the next trial still has exactly 5.56% odds — the universe doesn't owe you a catch-up.
This is the gambler's fallacy in action. The only real strategy is patience and volume. Keep incrementing your UUID counter, and the math will deliver.
[10]Start Your Collection
The buddy engine is deterministic, the math is known, and the tools are free. Whether you're aiming for a casual 18-species sweep or the insane 90-combo grid, the path is clear:
- Open the Buddy Checker and start entering strings
- Log every result in your zoo spreadsheet
- Track your progress against the milestone table
- Share your rarest finds using the share card feature
- Explore the Species Catalog to see what you're still missing
Remember: the median collector finishes all 18 species in just 59 trials. That's less than an hour with the Buddy Checker. Your zoo is waiting — go build it.