Seven runners finish a race in these times, in minutes: 31, 24, 28, 45, 26, 29, 27. Put them in order from fastest to slowest: 24, 26, 27, 28, 29, 31, 45. The runner in the middle position, the 4th, finished in 28 minutes. Three runners were faster and three were slower. That middle time is the median.
With an even number of values there is no single middle. Add an eighth runner at 30 minutes: 24, 26, 27, 28, 29, 30, 31, 45. Now the 4th and 5th values, 28 and 29, share the middle, and the median is halfway between them: (28+29)÷2=28.5. The position formula 2n+1 gives 4.5 here, which means halfway between the 4th and 5th values.
Go back to the seven runners and change the slowest time from 45 to 90 minutes, as if someone walked the course. The median does not move: the middle runner still finished in 28. The mean jumps from 30 to about 36.4. That resistance is the point of the median. But the median is not the middle of the range: the times run from 24 to 45, whose midpoint is 34.5, and nobody in the middle of the field ran anything like that.
Think of it like this. The median is the person standing in the middle of a line sorted by height, with as many people on one side as the other. Swap the tallest person for a basketball player and the middle person does not change. The picture needs one adjustment for an even-sized line, where the middle falls between two people and you take the value halfway between them.
Core idea. The median is the middle of the sorted data, with half the values on each side, so it is decided by position rather than by how large the extreme values are.