Dice Probability Basics for Board Gamers
You don't need to remember high-school math to play smarter board games. A handful of intuitions about how dice actually behave will change how you bid in Catan, when you take the safe move in Risk, and whether you go for it on a long-shot D&D attack.
One die is boring (and that's the point)
Roll a single fair six-sided die. Each face — 1, 2, 3, 4, 5, 6 — has the same chance: one in six, or about 16.7%. Roll a d20 and each of the twenty faces has a 5% chance. A d100 (a "percentile" die) gives you exactly 1% per face. This is called a uniform distribution: every outcome is equally likely. It's flat, predictable, and not very interesting on its own.
What makes dice interesting is what happens when you roll more than one and add them up.
Two six-sided dice: the bell curve
Roll 2d6 and the possible totals run from 2 to 12 — eleven outcomes total. But those outcomes are not equally likely. Here's the breakdown:
| Total | Ways to roll it | Probability |
|---|---|---|
| 2 | 1 (1+1) | 2.78% |
| 3 | 2 (1+2, 2+1) | 5.56% |
| 4 | 3 | 8.33% |
| 5 | 4 | 11.11% |
| 6 | 5 | 13.89% |
| 7 | 6 | 16.67% |
| 8 | 5 | 13.89% |
| 9 | 4 | 11.11% |
| 10 | 3 | 8.33% |
| 11 | 2 | 5.56% |
| 12 | 1 (6+6) | 2.78% |
Plot those percentages and you get a tidy little bell shape. Seven is the peak — there are six different combinations that add up to 7, while there's only one way to roll a 2 (snake eyes) or a 12 (boxcars). This is the single most useful fact in board gaming.
Why Catan punishes you for the 7
In Catan you place settlements on hexes labeled with numbers from 2 to 12. Each turn somebody rolls 2d6, and the hex with that number produces a resource. Beginners often grab a hex labeled "12" because the number looks high and important, then wonder why they're starving by turn 10.
A hex labeled 12 produces a resource only 2.78% of the time. A hex labeled 6 or 8 produces 13.89% of the time — roughly five times more often. A hex labeled 5 or 9 produces 11.11% — four times more often.
That's why Catan dots the high-frequency numbers (6, 8) with extra pips on the chits. Look for the 6s and 8s when placing settlements. Avoid the 2s, 3s, 11s, and 12s unless they're paired with a strong resource (like brick early game).
Then there's the 7. Rolling a 7 doesn't produce any resource — instead, the Robber blocks a hex and players with too many cards lose half their hand. Seven is also the most common roll, at 16.67%, so on average it lands every six turns. Plan accordingly.
What "average" actually means
The average roll of a single die is the midpoint between its lowest and highest face, plus one, divided by two. For a d6 that's (1+6)/2 = 3.5. For a d20 that's 10.5. For 2d6 it's 7 (which lines up with our bell curve).
Average matters because if a fight in D&D needs you to roll 12 or higher to hit, and your d20 averages 10.5, you'll miss more often than you hit unless you have a modifier. A "+4 to hit" turns the average from 10.5 to 14.5, swinging the encounter dramatically.
The d20: feast or famine
A single d20 is not a bell curve — it's flat. Every face is exactly 5% likely. That means swing rolls are common: a 1 happens 5% of the time, and so does a 20. This is why tabletop RPGs feel so dramatic compared to board games using 2d6 — every roll has genuine suspense, including the spectacular failure.
Two specific outcomes get extra weight by D&D convention: rolling a natural 20 (regardless of modifiers) is a "critical hit" — extra damage, drama, a moment for the table. Rolling a natural 1 is a "critical miss" — the opposite. Both happen 5% of the time. In a four-hour session with each player rolling 30 times, expect each player to crit roughly once or twice and fumble roughly once or twice. The numbers feel rare in the moment but are entirely predictable in aggregate.
Three or more dice: the curve gets sharper
Roll 3d6 and the totals run from 3 to 18, with the average at 10.5 and the curve much tighter than 2d6 — extreme outcomes (3 or 18) are now under half a percent each. The more dice you roll and add, the more the result clusters around the average. This is called the "central limit theorem," and it's why Yahtzee with 5d6 feels both surprising and predictable: any individual category is a long shot, but the total of all five is almost always near 17.5.
Practical takeaways
- On 2d6, the middle numbers (5–9) make up 67% of all rolls. The extremes (2, 3, 11, 12) are only 17% combined.
- On a d20, never feel cheated when you roll a 1 or 20 — they each happen 1 in 20 rolls.
- Adding modifiers shifts the average; even +1 changes your hit rate noticeably on a d20.
- The more dice you sum, the less variance — two-dice games swing a lot, five-dice games settle near the average.
Want to see this in action? Open the roller, pick the Yahtzee preset, and roll twenty times. You'll feel the bell curve under your fingers.
Related
Up next: D&D Advantage and Disadvantage: the math behind it.