How to Play Order Puzzles Without Guessing

Eight items can be placed in 40,320 different orders. If you try one order per minute and never repeat yourself, brute force takes about four weeks of continuous play. Nobody solves these games that way, and nobody is supposed to. Here is what to do instead.

Why is the order you choose really a time budget?

This is the one idea the whole genre rests on. After each placement, everything already on the board tries to advance one level. So the first thing you place gets a chance to grow on every turn that follows, and the last thing you place gets one chance, ever.

Turn that around and it becomes a rule you can actually use: look at how many levels each item needs, and give the long ones the early slots. An item with five stages cannot possibly finish if you place it sixth. You do not need to know anything else about the puzzle to rule that out, and ruling things out is most of the work.

How do you find the root of the dependency tree?

Every item's growth is gated by a condition, and those conditions point at other items. Write them out and you get a tree. Somewhere near the bottom of it there is an item whose own growth depends on nothing, or almost nothing, while several other items are waiting on it.

That item is the root, and roots go early. Not because it is obviously important, but because everything downstream of it is stuck until it moves. In most order puzzles there is exactly one, and finding it converts a wall of possibilities into a much smaller question: what goes second?

Read the chart, not the picture

It is tempting to place things because they look like they belong together. The art is deliberately charming and it will lead you astray. The evolution chart is the actual rulebook, and ten minutes with it beats an hour of pretty guesses.

Which items are hazards, and when should you place them?

Order puzzles usually include at least one item that damages its neighbours rather than cooperating with them. It will have a condition attached to the harm, something like "this hurts that, as long as that is still small."

Two things follow. First, the victim needs to outgrow the danger before the hazard arrives, so the hazard goes late. Second, and this is the part that catches people, the check usually runs every turn, not just on the turn you place it. A hazard sitting quietly on the board will still ruin something you add five turns later. If a run keeps dying to the same cause, look at what was already there rather than what you just placed.

Which items can you safely save for last?

Any item that only needs one or two levels is nearly free to place late. It does not need a long runway, so it can absorb whichever slot is left over. Identify those early and you have effectively shortened the puzzle: six real decisions instead of eight.

How should you experiment after a run fails?

Once you have a plausible order, do not throw it away when it fails. Change exactly one position and run it again. If you shuffle three items at once and the score moves, you have learned nothing about which change mattered.

Read the failure message every time. In our own hints section that is the only thing we tell you to do, because it is the only thing that generalises. A run that ends with six items maxed and one dead is not a failed run, it is a labelled data point: it tells you which pair is in conflict and roughly where in the sequence the conflict happens. Three or four careful runs usually produce more information than twenty careless ones.

What a good run feels like

You will know you are close when nothing dies and the score stops moving. That usually means the order is safe but somebody ran out of turns, which is a much smaller problem than a death: you are looking for two adjacent items to swap, not a new plan.

How many runs should it take?

Three to six careful runs is normal for an eight-item puzzle, if each run changes one thing. People who shuffle several items at once often spend thirty runs and end up knowing less.

The count matters less than what each run costs you. A run where you changed one position tells you what that position does. A run where you changed four tells you that something you did mattered, which you already knew. Keeping the number low is a side effect of keeping each run clean, not a target to aim at.

If you are past ten runs and still guessing, stop placing things and write the requirement list out instead. Most people who feel stuck have the information already and have never put it on paper, where a contradiction becomes obvious in a way it never is in memory.

What if two items both need to go first?

They cannot both go first, so one of them is not a true root. Check whether the second one actually needs the first, or only needs something the first one produces a turn later.

This is the most common deadlock in the genre and it is almost always a misreading. Requirements are usually stated as a level, not an existence: an item may need water at level two rather than water placed. That distinction changes everything, because it means the two can be two slots apart rather than competing for the same one.

When it is genuinely circular, the puzzle is telling you that one of the two is meant to finish below maximum. Some puzzles have no perfect ending at all; ours has exactly three orders out of 40,320 that reach 27, which means 40,317 of them leave something short. Knowing which item you are sacrificing is a plan. Discovering it on the last click is not.

Does this method work on other games?

Yes, on any puzzle where you choose a sequence and the board reacts after each choice. The item names change; the time budget, the dependency root and the hazard rule do not.

It transfers across the whole GROW family, from six-item entries up to the twelve-button ones, and to anything else built on the same skeleton. The larger the puzzle, the more it rewards the method rather than patience: at ten items there are 3,628,800 orders, and no amount of stubbornness covers that.

Where it stops working is games that add randomness or timing. If the same order can produce two different outcomes, runs stop being comparable, and the entire approach rests on runs being comparable.

No solutions here on purpose. The hints section on the game page is as far as we go.