Exactly 4 Gallons, or Else
Before you read any further, try it yourself. You’re allowed only three types of moves:
- Fill a jug completely from the fountain.
- Empty a jug completely.
- Pour from one jug into the other, until either the first jug is empty or the second jug is full, whichever happens first.
That’s it. No eyeballing “about half.” No guessing.
Take a minute. Grab a glass of water and an empty cup if it helps. See if you can land on exactly 4 gallons using only the 5-gallon and 3-gallon jugs.
Before you scroll past it, here’s why it’s worth the ninety seconds
This isn’t really a puzzle about water. It’s a puzzle about what happens when “close enough” isn’t an option.
Most of the time, we live in a world of “close enough.” A report that’s mostly accurate. A feature that’s mostly working. A task that’s mostly done. Nobody hands us a scale wired to a bomb, so we round up and move on.
But every so often, something forces a hard line. Exactly 4 gallons, not 3.9, not 4.1. And when that happens, the interesting part isn’t the water. It’s the fact that two people, McClane and Zeus, had to agree on what “exactly right” meant, agree on the rules for getting there, and trust that the other person was reading the jug the same way they were. One person guessing “close enough” would have gotten them both killed.
That same gap between “mostly done” and “actually done” is one that shows up constantly in how teams work together, and it’s the reason this puzzle is worth more than a minute of your time. Hold that thought. Try the puzzle first.
Still here? Let’s walk through it.
Step 1. Fill the 3-gallon jug completely.
5-gallon jug: 0. 3-gallon jug: 3.
Step 2. Pour all of it into the 5-gallon jug.
5-gallon jug: 3. 3-gallon jug: 0.
Step 3. Fill the 3-gallon jug again.
5-gallon jug: 3. 3-gallon jug: 3.
Step 4. Pour from the 3-gallon jug into the 5-gallon jug, but stop the moment the 5-gallon jug is full. It already had 3 gallons and can only take 2 more before it’s full.
5-gallon jug: 5. 3-gallon jug: 1 (because 3 minus the 2 that went in leaves 1).
Step 5. Empty the 5-gallon jug completely.
5-gallon jug: 0. 3-gallon jug: 1.
Step 6. Pour that 1 gallon into the 5-gallon jug.
5-gallon jug: 1. 3-gallon jug: 0.
Step 7. Fill the 3-gallon jug again.
5-gallon jug: 1. 3-gallon jug: 3.
Step 8. Pour it all into the 5-gallon jug.
5-gallon jug: 1 + 3 = 4. 3-gallon jug: 0.
Bomb defused.
Why this actually works
The reason this puzzle is solvable at all comes down to a small piece of number theory. Because 5 and 3 share no common factor other than 1, every whole number between 0 and 5 is reachable through some combination of fill, empty, and pour. If the jugs were 4 and 2 gallons instead, you’d never reach an odd number like 3, no matter how cleverly you poured. The constraint isn’t a limitation you work around. It’s the thing that makes a precise answer possible in the first place.
Teams often think a standard like this slows them down. In practice, it’s the opposite. Precision doesn’t come from having more freedom. It comes from having clear, fixed rules: fixed jug sizes, fixed moves, a fixed target. Take the fixed rules away and you don’t get more flexibility, you get a puzzle nobody can solve, because there’s no longer a shared basis for judging whether an answer is right.
What this has to do with Scrum
Here’s the part I keep coming back to.
The bomb didn’t accept “close enough.” 3.9 gallons: explosion. 4.1 gallons: explosion. There was no partial credit, no “we’re basically there,” no room for interpretation between McClane and Zeus about what “done” meant. Four gallons meant four gallons, and both of them had to agree on it before they touched the scale.
That’s the uncomfortable question a real Definition of Done is supposed to settle, and it’s exactly what most teams quietly skip.
Ask five people on a Scrum team what “done” means for a given item, and you’ll often get five slightly different answers. One person means “the code works on my machine.” Another means “the code works and there’s a test for it.” Another means “the code works, there’s a test, and someone else reviewed it.” Nobody’s lying. They’re just each carrying their own private version of “4 gallons,” and nobody wrote it down.
The jug puzzle has no ambiguity because the rules were fixed before anyone touched the water: fill, empty, pour until full or empty. A Definition of Done works the same way, with one important difference from the bomb. It doesn’t tell you what every single Product Backlog Item is supposed to do, the way “exactly 4 gallons” is one fixed number. It defines the minimum standard every piece of work must meet before it can honestly be called part of the Increment: things like “tested,” “reviewed,” “documented,” agreed on in advance, applied consistently. Individual items can carry their own extra acceptance criteria on top of that. But the floor itself doesn’t move, and like the puzzle, meeting it is binary. Either the work clears the standard, or it isn’t done. There’s no version of “80% done” that counts, the same way there’s no version of “3.9 gallons” that stops the bomb.
The upside is the same too. Once the rule is fixed and everyone can see the exact state of things, teams stop guessing and start reasoning. McClane and Zeus didn’t need more water or a bigger jug. They needed clarity on what counted as the target, and they needed to know exactly what was in each jug before deciding their next move. Three gallons here, one gallon there, and the next step followed from that, not from a hopeful guess. That’s empirical process control in miniature: transparency about the real state of things is what makes honest inspection possible, and honest inspection is what gives you something real to adapt to.
Next time your team debates whether something is “basically done,” it might be worth asking: do we actually have a Definition of Done, or are we just hoping the bomb won’t notice?
