The Three-Level Puzzle Challenge
Try seven puzzle questions and inspect how the original score is calculated.
About the archived wording: percentage, rarity, IQ, timed-performance and mind-reading claims in the original copy below are fictional challenge language, not measured findings. The rules and result labels are preserved for entertainment.
Level one lets almost everyone through. It smiles, it flatters, it hands you an easy win to keep you walking. Level two starts sharpening its questions. And level three is where the confident tend to go very quiet.
This is a gauntlet of trick questions wearing the costume of simple ones, the sort that punish the fast and reward the ones who pause. There is no clock and no penalty for thinking twice, which is exactly the trap. Read every single word, assume nothing, and watch for the question that is quietly laughing at you. Clear all three levels and you join a very small club.
Read every possible result and what it means →
Inspect the scoring, question by question
The highest possible score here is 21. The original game divides scores into 4 bands; the bands contain different numbers of answer combinations. A high point total is not a measured percentile.
1. Level one, nice and gentle: a farmer keeps some sheep, and all but four of them wander off. How many sheep does he have left?
- Zero; they all wander off eventually — 1 original points
- Four — 3 original points
- You can't tell without the starting number — 2 original points
- Sheep don't wander, so the question is broken — 0 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
2. Still warming up: you are in a footrace and you overtake the runner sitting in second place. What position are you in now?
- First — 1 original points
- Third — 0 original points
- Second — 3 original points
- Impossible to say without more detail — 2 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
3. Level two, a little sharper: some months have thirty days, some have thirty-one. How many months of the year have twenty-eight days?
- Just one, February — 1 original points
- Four — 0 original points
- It depends on leap years — 2 original points
- All twelve of them — 3 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
4. Weigh it carefully: which is heavier, a pound of feathers or a pound of bricks?
- They weigh exactly the same — 3 original points
- The bricks, obviously — 1 original points
- The feathers — 0 original points
- It depends on the size of each pile — 2 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
5. Getting mean now: a rooster perches on the very peak of a slanted barn roof and lays an egg. Which side of the roof does the egg roll down?
- Down the steeper side — 1 original points
- Whichever side happens to face downhill — 0 original points
- Down the side facing into the wind — 2 original points
- Neither; roosters don't lay eggs — 3 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
6. Quick arithmetic, or is it: how many times can you subtract five from twenty-five?
- Five times — 1 original points
- Twenty-five times — 0 original points
- Once; after that you're subtracting from twenty — 3 original points
- Infinitely, as often as you like — 2 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
7. The level three boss: before the tallest mountain on Earth was ever discovered and measured, which mountain was the tallest?
- K2 held the title back then — 1 original points
- There simply wasn't one yet — 0 original points
- The same peak; it was tallest all along, just undiscovered — 3 original points
- There is no way to ever know that — 2 original points
These are the archived game’s option weights. Partial credit is a game rule, not a clinical or population measurement.
What happens across every possible answer pattern?
We counted all 16,384 patterns, treating each offered choice as equally likely. This is a synthetic rule check, not visitor research. It does not describe how real people answer.
| Result | Patterns | Share of patterns |
|---|---|---|
| Wiped Out at Level 1 | 323 | 1.97% |
| Tripped Near the Top | 7,869 | 48.03% |
| One Wall to Go | 7,456 | 45.51% |
| Broke the Level-3 Curse | 736 | 4.49% |
Download the full counts · Download Python script · Download rule inputs
To reproduce locally: python3 measure_legacy.py --input legacy-rule-inputs.json --output counts.json. Uses Python 3 and its standard library.
Arcade difficulty was an accounting decision
Inside a coin-operated cabinet, difficulty is a pricing mechanism. Operators wanted each paid turn to last a fairly predictable few minutes so that a machine earned steadily through a busy Saturday, and designers tuned the ramp to suit: a generous opening to hook somebody walking past, then a steepening curve that dependably ends the session. Extra lives, continues and high-score tables all grew out of that arrangement.
Home consoles inherited the shape without the economics. A cartridge bought once had no reason to end anybody's turn, but the games were short and had to feel worth the shelf price, so difficulty stayed brutal as a way of stretching them. Players later gave that era the affectionate label of Nintendo hard.
The most famous wall in the medium
Ask a certain generation about level three and one answer comes back: the Turbo Tunnel, third stage of Rare's Battletoads on the NES in 1991. It abandons brawling for a speeder bike that accelerates into a run of walls requiring memorised, frame-precise jumps, with no checkpoint worth the name. It is a genuine skill gate, the point at which a game stops asking whether you are enjoying yourself and starts asking whether you have practised.
Skill gates cut both ways. They give the eventual mastery moment real weight, and they quietly end the experience for most of the people who ever loaded the cartridge. Modern designers tend to soften them with generous checkpoints, difficulty options and practice modes, which is why an entire genre could be built later on the promise of putting the wall back in and making the wall the point.
Trick questions are gates as well
The gauntlet above pulls a different lever. Its walls are not reflex tests but readings: questions whose obvious answer has been deliberately planted, so speed gets punished and a second pass gets rewarded. That is a design choice about attention rather than skill, which is why confident readers frequently do worse than hesitant ones.
Treat a modest tier as evidence that you read at a normal human pace, which is, most days, exactly the correct way to read.
Reading context & limits
For background and a fuller source trail, read Why Personality Quiz Results Feel So Accurate. These references provide context; they do not validate the entertainment result or every historical detail in the archived essay.
Content lead: LISA. Updated October 25, 2026. Send a specific factual correction.