Most advice about hidden-piece games like Stratego is vague hand-waving about "reading your opponent". This page is the opposite: a solo puzzle has exact numbers, so here are the exact numbers, for the twelve-piece board on the home page.
The one fact everything follows from
The flag is always in the back row. Every strategy on this page is a consequence of that sentence. The front row holds three pairs (Scout, Sapper, Lieutenant) that can never be the flag and are never adjacent to it. So an order spent on the front row reveals something true, and useless.
That gives the first rule: spend nothing on the front row, ever. Three orders all go into six squares.
The baseline you are trying to beat
Pointing at the flag with no orders wins one time in six: 16.7%. It also pays 133 points, the highest tier in the game. So the real question is not "how do I win", it is "how many orders are worth spending before a win stops being worth much".
What each order actually buys you
The back row is not six independent squares. It is three adjacent pairs - Captain and Captain, Colonel and Colonel, and Marshal sitting next to Flag - in some order. That pairing is what makes a single order informative.
| First order reveals | How often | Resulting position |
|---|---|---|
| The Flag itself | 1 in 6 | Round over, you have won |
| The Marshal | 1 in 6 | Flag is one of its two neighbours, so 1 of 2 |
| A Captain or a Colonel | 4 in 6 | That pair is ruled out, Flag is in one of the other two pairs, so 1 of 4 |
Combine those and one order takes you from 16.7% to about 41.7%: one sixth of the time you win outright, one sixth of the time you are down to a coin flip, and two thirds of the time you are down to one in four.
That is why one order is the biggest single jump in the game. It more than doubles your chance.
The Marshal case is already solved
When you flip the Marshal, do not think of it as a coin flip. You still have orders left, and the Flag is one of exactly two neighbours. Flip either one of them:
- If it is the Flag, you have won.
- If it is not, then the Flag is the other neighbour, and you have won anyway.
So the Marshal case needs two orders in total (the one that found it plus one more) and resolves to certain victory. One sixth of all rounds are free wins for anyone who knows this.
The two-order rule
Two orders is where this puzzle stops being obvious. Go through it case by case.
If the first order found the Flag (1 in 6) or the Marshal (1 in 6), you are already certain. That is a third of all rounds done in at most two orders.
The other two thirds of the time you have revealed a Captain or Colonel pair. You now know the Marshal-and-Flag pair occupies one of the remaining two pair positions, four squares in all. Your second order goes into one square of one of those two pairs:
- It comes up Flag - a quarter of the time from here - and you have won.
- It comes up Marshal - another quarter - so the Flag is the Marshal's neighbour, and you have won.
- It comes up a Captain or Colonel - the remaining half - which means the Flag is one of the two squares in the last untouched pair, and you point with a 50-50 shot.
Add that up and two orders gets you to about 83.3%.
A third order takes you to essentially 100%, because worst case you are choosing between two squares and one order tells you which.
So how many should you spend?
Win rate is only half the story; the scoring tiers are the other half. Multiply one by the other and you get expected points per round:
| Orders spent | Win chance | Points if right | Expected points |
|---|---|---|---|
| 0 | 16.7% | 133 | about 22 |
| 1 | 41.7% | 100 | about 42 |
| 2 | 83.3% | 67 | about 56 |
| 3 | about 100% | 33 | about 33 |
The mathematically best line is exactly two orders. Not zero, not three. Three orders wins almost every round and pays so little that it is worse than spending half as much and missing occasionally. That is the tradeoff the scoring is designed to force, and it is why a blind call still sits at the table.
Two caveats worth being honest about: these figures describe optimal play on this specific twelve-piece board, with a fixed pair structure and the Flag known to be in the back row. Change any of those and the numbers change. And a streak rewards a win more than a high score, so if you are chasing consecutive days rather than points, spending all three is the defensible choice.
Worked example
You have three orders. You spend none and you point at random: that is the 16.7% line, worth 133 if it lands.
Instead, spend one into the back row and it comes up Colonel. The Captain pair and the Marshal-Flag pair now occupy the other two pair positions. Spend your second order into one of those four squares: three times out of four you either find the Flag, find the Marshal, or get told the Flag is next door. Only in the remaining case do you spend the third order to break a two-way tie. Two orders used, 67 points, one in five rounds you get unlucky.
The five mistakes that cost most
- Scouting the front row. Three free pieces of information about pieces that cannot help you find the Flag.
- Treating the Marshal reveal as a coin flip. It is not, it is solvable with one more order.
- Spending the third order reflexively. With two orders spent you are usually at 50-50 on two squares. Spending the third turns a 67-point win into a 33-point one, which is a bad trade if the tie is genuinely even.
- Memorising positions. The formation is different every day. The method transfers, the board does not.
- Forgetting the pairs. Captain and Captain are adjacent, Colonel and Colonel are adjacent, Marshal and Flag are adjacent. Half of all deduction on this board comes from that single rule.
Frequently asked
Is two orders really better than three?
On points, yes: about 56 expected points against about 33. Three orders maximises your chance of winning the round, two orders maximises your score. Pick whichever you are actually optimising for.
Should I ever point with zero orders?
Yes, in the same sense that a poker player sometimes shoves on a bluff: the payout is 133, four times a fully-informed win. Do it rarely, and only when you accept the 16.7%.
Does the strategy change day to day?
The structure does not: pairs, Flag in the back row, three orders. Only the arrangement changes, so the priority order - back row first, Marshal before guessing, two orders unless you are chasing a streak - holds every day.