Understanding histograms, one trial at a time
Every time something happens, you get a number out of it: how long a card took, how many bugs showed up, what a die landed on. A histogram just counts how often each number showed up, and stacks those counts into bars. Once you've seen a few, the shape of the bars starts telling you something on its own.
Flip a fair coin and you'd expect heads and tails to come up about equally often. Flip it a few times yourself, then flip a bunch more at once, and watch the two bars settle out.
Given enough flips, the two bars end up close to even. Neither side is "due" for a turn: a coin has no memory of its last flip.
A die has six sides instead of two, but the same idea holds. Give it enough rolls and every face should show up about as often as the rest.
Same flat shape as the coin, just spread across six bars instead of two.
If you enjoy tabletop games, this histogram lets you do some forecasting. Let's say you're playing a game where you need to roll at least a 3.
The forecast is only as good as the sample size it's based on: a handful of rolls can swing it wildly, but the more you roll, the closer it settles toward the true odds.
Now roll two dice at once and add them together. Each roll gives one combined number, anywhere from 2 to 12.
This one doesn't look flat like the other two. The numbers in the middle show up more often than the ones at either end.
You can turn that into an actual number. Pick a value and see how likely it is, based on what you've rolled so far.
7 is the single most common roll here. But "most common" doesn't mean "likely": set the picker above to exactly 7 and you'll see the odds are still only around one in six. Plan around 7 as if it were a sure thing, and you'd be wrong five rolls out of six.