Higher or Lower
Guess whether the next card will be higher or lower and build a streak — with the counting that turns a guessing game into a calculable one.
Will the next card be higher or lower?
An Ace is a guaranteed correct answer
A match counts in your favour here, and that one rule turns two cards into free points. From an Ace, "higher" wins on all 13 outcomes: twelve are genuinely higher and the thirteenth is a tie. A King and "lower" is the same. Nothing else in the deck is certain.
| Showing | Say | Chance | Expected streak from here |
|---|---|---|---|
| A | higher | 13/13 · 100.0% | 4.27 |
| 2 | higher | 12/13 · 92.3% | 3.87 |
| 3 | higher | 11/13 · 84.6% | 3.49 |
| 4 | higher | 10/13 · 76.9% | 3.15 |
| 5 | higher | 9/13 · 69.2% | 2.83 |
| 6 | higher | 8/13 · 61.5% | 2.53 |
| 7 | either | 7/13 · 53.8% | 2.26 |
| 8 | lower | 8/13 · 61.5% | 2.53 |
| 9 | lower | 9/13 · 69.2% | 2.83 |
| 10 | lower | 10/13 · 76.9% | 3.15 |
| J | lower | 11/13 · 84.6% | 3.49 |
| Q | lower | 12/13 · 92.3% | 3.87 |
| K | lower | 13/13 · 100.0% | 4.27 |
The rule is a clean threshold: higher below seven, lower above seven. On a seven the two options are identical at 7/13, so it is both the worst card in the deck and the only one where no skill is available. Everything above is a mirror of everything below it. Play the threshold every time and you win 78.7% of individual guesses.
Guessing at random is exactly as good as holding a 7
Pick a button without looking and your chance is the average of the two options, which works out to 7/13 — 53.8% — no matter which card is showing. The card genuinely does not matter if you are not using it. And 53.8% is precisely what an expert gets on a seven. Always pressing "higher" comes to the same average too: the low cards it wins on and the high cards it loses on cancel exactly.
So the entire skill in this game is the gap between 53.8% and 78.7% — 24.9 percentage points — and all of it lives in the extreme cards. There is no clever play to find on the middle ones; there is only the discipline of not talking yourself out of the obvious one.
Perfect play still only averages a streak of three
Winning 78.7% of guesses sounds like long runs. Solving the chain exactly gives an expected streak of 3.27 under perfect play — 4.27 if you open on an Ace or a King, 2.26 if you open on a seven.
That is lower than the 3.69 a flat 78.7% win rate would suggest, and the reason is worth a moment: surviving a guess tends to leave you holding a middling card. The good cards are good precisely because they send you almost anywhere, and almost anywhere is usually somewhere worse than where you were.
Without the tie rule the same perfect play would average 2.23. Counting a match as a win is worth 47% on your streak, and it helps from every card, not just the two extremes. It is a large thumb on the scale for a rule most players never notice is there.
One last thing the maths depends on: every card here is drawn fresh from thirteen ranks, so there is no deck to run down and counting is useless. In a real 52-card deck the next card would be one of 51, of which 3 match — a 5.9% tie against 7.7% here. This game is a little more generous than the cards it draws pictures of.
How to use
- Look at the current card.
- Guess whether the next is higher or lower.
- Build your streak without a wrong call.
- Track which cards have already gone.
Frequently asked questions
Is there any skill in this?
More than it appears. With a single deck that is not reshuffled, every card dealt changes the odds on the next, so tracking what has gone turns guesswork into arithmetic. On a middling card the call is close; on an extreme one it is nearly certain.
What is the correct call on a 7?
In a fresh deck it is almost exactly even, since a 7 sits at the middle of thirteen ranks — which makes it the hardest card in the game. This is why a run of middle cards is where streaks tend to end.
How much does counting help?
Substantially, in a game without reshuffling. If most low cards have already appeared, higher becomes far less attractive than the face value suggests. Even a rough sense of which half of the deck has been depleted improves the decision noticeably.
What happens on a tie?
It depends on the rules in force, and it is worth knowing which apply — some count a tie as a loss, some as a win, some deal again. With four cards of each rank the situation arises often enough to matter to your strategy.
Why do streaks end so abruptly?
Because the probability of a long run is the product of each individual call, and even good calls at 70 per cent compound quickly downward. Ten correct calls at that rate happens under 3 per cent of the time, which is why long streaks feel rare — they are.
Is this the same as the pub machine version?
The mechanic is, and those machines are worth understanding as a contrast: they typically reshuffle or use odds that make sustained play unprofitable regardless of skill. Here there is nothing at stake, which is the material difference.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.