What a published rate tells you
Many games that sell randomised rewards now publish the chances of what a box can contain. That is helpful, but a table of percentages is easy to misread. The numbers describe a single open under particular rules, and several details determine what they mean for the thing you actually want.
This article is educational. It explains how to read the figures and does not suggest that any approach makes outcomes more favourable. Purchases of random rewards carry real uncertainty and cost, and rules about disclosure and age limits vary by region.
Tier odds are not item odds
A common layout lists chances by rarity: for example common 70 percent, rare 25 percent, epic 4 percent, legendary 1 percent. These add to 100 percent. The trap is that the legendary tier may contain many items.
If that 1 percent legendary chance is shared evenly between five legendary items, then each has a 0.2 percent chance per open (1 divided by 5 is 0.2). The expected number of opens to get one specific legendary is 1 / 0.002 = 500, five times the figure you would get from reading the tier rate alone. If the table does not say whether items inside a tier are equally likely, that is a question worth asking before spending anything. Some tiers use unequal weights, as described in the article on drop tables.
Chance across several opens
Because each open is typically independent, the chance of at least one hit across n opens is 1 minus the chance of missing every time. Take a featured item with a 2 percent chance per open. The chance of missing in one open is 0.98, so the chance of missing in ten opens is 0.98^10, about 0.817. The chance of at least one hit is therefore about 18 percent, not the 20 percent that simple multiplication (10 times 2) suggests.
That small gap grows with more opens, and it points to an important idea: chances do not add up past a point. Ten opens at 2 percent do not make 20 percent, and 50 opens do not make 100 percent. At 50 opens the chance of at least one hit is 1 minus 0.98^50, about 64 percent. The remaining 36 percent of buyers would still be without the item.
The average cost is the reciprocal of the chance: 1 / 0.02 = 50 opens. Half of buyers need fewer than that, some need many more, and the spread is wide, as explained in the article on expected attempts.
Bundled and guaranteed pulls
Multi-open offers often add a guarantee, such as “at least one rare or better in every ten”. This changes the arithmetic in a specific way. If the base rare-or-better chance is 5 percent per open, then without the guarantee the chance of getting none in ten is 0.95^10, about 60 percent. With the guarantee, none is impossible in that bundle.
But the guarantee only covers the tier named. It says nothing about whether you receive the particular item you want, and the guaranteed slot may draw from a separate, less generous table. Read the fine print on which slot has which odds.
Duplicates change the real value
Even if you do receive rare items, duplicates reduce the value of each. Suppose a rare tier holds four equally likely items and you want a complete set. The average number of rare drops needed to collect all four is 4 times (1 + 1/2 + 1/3 + 1/4), which is about 8.3. If rares arrive at 5 percent per open, that is around 8.3 divided by 0.05, or roughly 167 opens on average.
Some games soften this with duplicate protection or conversion into currency. Whether that is present, and how generous the conversion is, can matter more to your real cost than the headline rate.
Things that quietly alter the numbers
- Rates that differ by pack, season or banner. The table you read may not apply to the box you open.
- Pity or guarantee counters. Counters that raise or guarantee a reward after repeated misses change the average, as the article on pity timers explains. Make sure the published rate is stated as the base rate or as the effective rate.
- Rounded figures. A listed 0.5 percent may be a rounded value. Over many opens, small differences add up.
- Odds that are not published. If only some tiers are disclosed, the rest cannot be assessed.
Why design and trust come into it
Randomised rewards sit at the intersection of enjoyment, spending and fairness. Clear disclosure helps players judge what a purchase is worth, and it is a recurring theme in monetisation design. Clear numbers do not remove the uncertainty of any single open, though. They give you the means to understand what the typical, and the unlucky, experience looks like before you decide.
A short reading checklist
- Is the rate per open, per pack, or per bundle?
- Is a tier chance split across items, and equally?
- How many opens would it take to reach, say, a 90 percent chance of your target? For 2 percent, 1 minus 0.98^n reaches 90 percent at about 114 opens.
- Are there duplicates, and what do they convert into?
- Do any counters or guarantees apply, and how are they described?
Working through these questions turns a table of percentages into a realistic picture of cost and variation, which is the point of publishing odds at all.