A tabletop damage calculator computes exact probability distributions for multi-step attack sequences in Warhammer 40K, Age of Sigmar, and similar games. Enter attacks, hit threshold, wound threshold, save, and damage to see the expected damage, the 10th and 90th percentile outcomes, and the full distribution behind them.
Stage-by-Stage Breakdown
| Stage | Input | Prob Each | Expected Out | 10th %ile | 90th %ile |
|---|
Wargaming Dice Gear
If you're rolling 40 attacks at once, standard board game dice won't cut it. To execute your hyper-optimized strategies efficiently, you need a high-quality bulk dice set. We recommend grabbing a brick of 36 Chessex 12mm D6s. Their smaller size allows you to roll massive handfuls without scattering them off the table. Furthermore, to protect your models and roll quietly during long tournament days, always use a dedicated neoprene dice tray.
How to Calculate Tabletop Damage
In games like Warhammer 40,000, calculating expected damage isn't as simple as multiplying the number of attacks by the flat probability of each roll. If a squad shoots 20 times, hitting on 3+, wounding on 4+, against a 5+ armor save, the outcome is a multi-step "attrition waterfall." Because dice rolls are discrete events, the variance at the beginning of the attack sequence drastically affects the end result. DiceDecay doesn't just give you an "average" damage number; it convolves the whole sequence into one exact probability distribution. This reveals the actual shape of the bell curve, allowing you to see your 90th percentile "Spike" damage, and more importantly, your 10th percentile "Guarantee" damage floor. Stop relying on averages and start playing the odds you can actually count on.
Frequently Asked Questions
Does this use binomial or hypergeometric probability?
It is binomial, and for dice that is the correct model. Each die is an independent roll with a fixed success chance, so the count of successes at each gate is binomial, and chaining the gates gives an exact compound binomial: thinning Binomial(N, hit) by the wound chance gives Binomial(N, hit × wound), and thinning that by the failed-save chance gives Binomial(N, hit × wound × save-fail). The hypergeometric distribution describes drawing from a finite pool without replacement, which does not apply to dice - rolling a 6 does not remove the 6 from the die. DiceDecay convolves the whole pipeline rather than multiplying averages, so it reports the full distribution over Attacks → Hits → Wounds → Unsaved Damage, not just the mean.
What does the 'Gamble' vs 'Guarantee' metric mean?
Gamble is the mean: the damage this attack averages over many repetitions, and the number the dashed line marks on the distribution chart. Guarantee is the 10th percentile, the floor you meet or beat nine times in ten when the dice run cold. Spike is the 90th percentile, the ceiling you reach one time in ten when they run hot. The gap between Guarantee and Spike is how swingy the unit is - low-attack, high-quality units have narrow ranges, and a wide range means the average is a number you will rarely actually roll.
How do reroll mechanics affect expected damage?
Rerolling 1s raises the success chance by exactly one sixth at every threshold - a flat +16.7%. One die in six is rerolled, so the reroll adds p/6 successes on top of p, giving p × 7/6 whatever the threshold. It is not one sixth of the failures that convert: that share is (p/6)/(1 - p), which runs from 83% on a 2+ down to 3% on a 6+ and only happens to equal one sixth on a 4+. Full rerolls (reroll every miss) give p(2 - p) instead, worth +33% on a 3+ and +50% on a 4+, because a dice can only be rerolled once. The gain carries through the pipeline unchanged - 33% more hits is 33% more wounds and 33% more unsaved damage.