Floor and Ceiling Advanced. This lesson goes beyond core Algebra I. You can skip it.

What the floor and ceiling are

The floor of $x$, written $\lfloor x \rfloor$, is the greatest integer that is less than or equal to $x$. The ceiling of $x$, written $\lceil x \rceil$, is the least integer that is greater than or equal to $x$. The bent brackets are a memory aid: the floor’s brackets point down toward the ground, the ceiling’s point up.

Try a few. Since $3.7$ sits between $3$ and $4$, the integers at or below it are $3, 2, 1, \dots$, and the greatest is $3$, so $\lfloor 3.7 \rfloor = 3$. The integers at or above $3.7$ are $4, 5, 6, \dots$, and the least is $4$, so $\lceil 3.7 \rceil = 4$. When the input is already an integer, like $5$, it is itself the greatest integer at or below and the least integer at or above, so $\lfloor 5 \rfloor = \lceil 5 \rceil = 5$.

Every one of those computations used the same test: find the integer that fits snugly against $x$ from below (for the floor) or from above (for the ceiling). That test is worth writing as an inequality, because it is the engine behind everything else.

Only one integer can do this. Consecutive integers are always a whole unit apart, so no two of them can both sit at or below $x$ while staying within one unit of it. That is what makes the inequality a computation tool: find any integer $k$ with $k \le x < k + 1$, and it has to be $\lfloor x \rfloor$.

The same reasoning, applied to the least integer at or above $x$, gives the mirror statement for the ceiling:

These two inequalities also settle what happens at the boundary between the two functions. When $x$ is an integer, it is both the greatest integer at or below and the least integer at or above itself, so

When $x$ is not an integer, the left inequality $n \le x$ becomes strict, $n < x < n + 1$, so $x$ lands strictly between the consecutive integers $n$ and $n + 1$. The least integer at or above $x$ is then $n + 1$, which gives a clean relationship between the two:

So the ceiling and floor are equal exactly on the integers, and exactly one apart everywhere else.

Rounding down is not the same as chopping off the decimal

Everything above is easy to misread once negative numbers appear, and this is the single most common mistake with these functions. “Round down” means move to the left on the number line, toward smaller values. It does not mean “delete the part after the decimal point.”

Take $x = -2.3$. It sits between $-3$ and $-2$. The integers at or below $-2.3$ are $-3, -4, -5, \dots$, and the greatest of them is $-3$, so

The tempting answer $-2$ is wrong, because $-2$ is greater than $-2.3$, so $-2$ is not “at or below” $-2.3$ at all. Chopping the decimal off $-2.3$ would give $-2$, but that is rounding toward zero, not rounding down. The floor always rounds toward negative infinity. The ceiling goes the other way: the least integer at or above $-2.3$ is $-2$, so $\lceil -2.3 \rceil = -2$. On the picture below, the floor is simply the tick to the left and the ceiling is the tick to the right.

The graphs are staircases

To graph $y = \lfloor x \rfloor$, group the inputs by the value they produce. The defining inequality says $\lfloor x \rfloor = n$ exactly when $n \le x < n + 1$. So on the whole interval from $n$ up to (but not including) $n + 1$, the floor holds steady at the height $n$. That makes the graph a flat horizontal segment at height $n$ running across that interval. The graph then jumps up by one at $x = n + 1$ to start the next segment. The result is a staircase.

Look carefully at the ends of one step. On the interval for height $n$, the left end $x = n$ is included, because $\lfloor n \rfloor = n$, so that end is a filled (closed) dot. The right end $x = n + 1$ is not included, because at $x = n + 1$ the floor has already jumped to $n + 1$, so that end is a hollow (open) dot. Every step of the floor is therefore closed on the left and open on the right.

The ceiling graph is the same staircase read the other way. Since $\lceil x \rceil = n$ exactly when $n - 1 < x \le n$, each ceiling step holds the height $n$ across the interval that ends at $x = n$. For a ceiling step the included end is now on the right (because $\lceil n \rceil = n$), and the excluded end is on the left. So every ceiling step is open on the left and closed on the right, the mirror image of the floor.

Both graphs pass the vertical line test (each input has one output), so they really are functions. Both accept every real number, so the domain of each is all real numbers. And every output is an integer, with each integer height actually reached, so the range of each is the integers. The steps never connect with vertical lines, because the function does not take the in-between values during a jump. The open and closed dots are what carry that information.

The fractional part

Subtracting the floor from a number strips away the whole part and leaves only what is left over. That leftover is the fractional part of $x$, written $\{x\}$:

It is always at least $0$ and always less than $1$, and the defining inequality shows exactly why. Start from $n \le x < n + 1$ with $n = \lfloor x \rfloor$ and subtract $n$ from every part:

For a positive number this matches intuition: $\{3.7\} = 3.7 - 3 = 0.7$, the digits after the decimal point. But watch what happens with a negative input, where the floor surprise strikes again. Because $\lfloor -2.3 \rfloor = -3$,

The fractional part of $-2.3$ is $0.7$, not $0.3$. It has to be, since $\{x\}$ can never be negative, and it measures how far $x$ has climbed above the step below it. On the integers the leftover is nothing at all, $\{5\} = 5 - 5 = 0$, which is the only time the fractional part equals $0$.

Where floor and ceiling show up

The reason these functions matter is counting. Whenever a real-number answer has to be turned into a whole number of things, one of the two brackets is exactly the right tool. The choice between the two brackets is never arbitrary.

Use a ceiling when every leftover still needs a full container. If $53$ students travel in vans that seat $12$, then $53 \div 12 = 4.41\ldots$ vans’ worth of seats are needed. But you cannot send a fractional van, and the $5$ students left over after four full vans still need one more. So the number of vans is $\lceil 53 / 12 \rceil = 5$. Buses for a crowd, pages for a list, and boxes for a shipment all work this way.

Use a floor when leftovers are discarded and you only count what is complete. From those same $53$ students you can fill $\lfloor 53 / 12 \rfloor = 4$ vans completely. The amount left over is what the full groups do not use,

which is the same fractional-part idea from before, now scaled up by the group size. Full crates from a harvest, whole dollars from a pile of change, and complete laps around a track are all floors.

Both answers live in one picture. Lay $n$ objects into rows that hold $c$ each: the number of complete rows is $\lfloor n/c \rfloor$, and the number of rows needed to hold everything, counting a partly filled last row, is $\lceil n/c \rceil$. Try it below.

Start at $10$ dots in rows of $4$. Two rows fill and two dots are left hollow, so $\lfloor 10/4 \rfloor = 2$ while $\lceil 10/4 \rceil = 3$. Any leftover from one dot to three would force that same extra row, since the row width is four. Now widen the rows to $5$: the same $10$ dots fill two rows exactly with nothing hollow, so floor and ceiling agree at $2$. Narrow the rows to $3$ instead and one dot is hollow again, splitting the answers to $3$ and $4$. Change the row width a few more times and watch for the pattern: the two functions differ by exactly one whenever a dot is left hollow, and are equal whenever none is.

Common mistakes