A dice roll doesn’t care how a player feels about it. That’s the whole problem, really. Every random encounter table, loot drop, or crit chance is math running quietly in the background, and players rarely see it directly. They just feel the results. Five bad rolls in a row and suddenly the game feels broken, even if the math behind it was working exactly as intended the whole time.
When fair math feels unfair
Here’s the thing about encounter systems: pure randomness often reads as unfair even when it’s technically correct. A coin-flip mechanic with no memory, no adjustment, no hidden weighting can produce a nasty losing campaign purely by chance, and there’s nothing wrong with the code when that happens. Players in Canada, in Brazil, anywhere at all, tend to read that streak as evidence the system is rigged against them specifically. It isn’t. It’s just a small sample doing what small samples do, which is occasionally look nothing like the long-run average a designer had in mind when they set the numbers.
Most designers who’ve been burned by this learn to bend the randomness a little, usually after watching enough players complain about the same thing. Pity timers show up a lot, quietly incrementing a counter until a guaranteed good outcome finally lands. Soft caps after a few bad results in a row are common too.
Small sample sizes and the arguments they cause
Most players never rack up enough rolls to see a stated probability actually play out the way the spreadsheet promised. A few dozen encounters over a single playthrough, maybe, not the thousands it would take for a five percent drop rate to reliably look anything like five percent. That gap explains a lot of the arguments that show up in patch note comments and player reviews, where someone insists a mechanic is broken because they got unlucky twice in a row and assumed that meant something.
Good designers tend to work around this by building feedback into the moment itself rather than just trusting players to do the statistics in their head, which almost nobody actually does well under pressure. A rough indicator before the roll helps. So does a brief explanation after an unusual outcome, something as simple as flagging that a streak was statistically unlikely rather than letting players draw their own, often wrong, conclusions.
What the casino industry already figured out
Casino games have dealt with this exact tension for decades, partly because regulators force a level of disclosure most video games never have to bother with. Anyone curious what that transparency actually looks like in practice can dig through the detailed breakdowns published on sites such as Casino.org, where return-to-player figures and volatility ratings get laid out game by game, making the underlying probabilities easier to understand rather than burying them in a terms and conditions page nobody reads. Nobody’s suggesting a fantasy RPG needs a regulator standing over its shoulder checking every loot table.
Games that hide their true probabilities behind soft, evasive language tend to lose player trust faster than games that just state the number plainly, even when that number isn’t especially flattering to look at. Players seem to forgive bad odds more easily than they forgive discovering, usually through a community spreadsheet someone built out of frustration, that they’d been kept in the dark about them the whole time.
Difficulty tuning is probability tuning, whether designers admit it or not
It’s easy to treat difficulty and probability as two separate design problems that happen to sit near each other. They’re not, not really. Shifting probability across an encounter table changes how often a player runs into a brutal fight versus an easy one, and that shapes perceived difficulty just as much as any stat block adjustment would, arguably more, since players rarely notice a five percent damage buff on an enemy but absolutely notice facing three hard fights in a row.
This matters most in games where threat is supposed to ramp up gradually over the course of a session or campaign. Get the curve too flat, and encounters start feeling repetitive fast, the same handful of enemy types cycling through with nothing new to react to. Too steep, and players hit a wall they don’t have the resources to deal with yet, usually right after a section that felt manageable, which makes the spike feel arbitrary rather than earned.
Showing the numbers without killing the tension
There’s a genuine tradeoff between transparency and mystery here, and it doesn’t resolve cleanly. Print the exact odds next to every possible outcome, and there’s a real risk of turning what should feel uncertain into something closer to a spreadsheet exercise, where the “correct” decision practically makes itself and any sense of risk evaporates. Some designers dodge this by showing relative risk instead of a precise percentage: high, medium, low, which keeps players reasonably informed without stripping away the tension that made the encounter worth engaging with to begin with.