A recipe promises 24 muffins. Some batter clings to the bowl, one muffin sticks to the tin and breaks, and you end up with 21 whole muffins on the rack. The 24 is the prediction; the 21 is what you can count. In chemistry the prediction is the theoretical yield, worked out from the limiting reactant, and the count is the actual yield: the mass of product you collect, dry and weigh.
Real reactions lose product in predictable ways. Some stays dissolved in the liquid poured off after filtering. Some sticks to glassware or filter paper. Some reactant never reacts, or turns into a side product instead. That is why an actual yield of 4.1 g against a theoretical 5.0 g is normal, and why actual yield has to be measured: no formula can predict how much your particular flask lost.
An actual yield above the theoretical yield is a warning sign, not a success. It usually means the product is still wet with solvent or mixed with leftover reactant, both of which add mass. If a dried precipitate weighs 5.6 g when the equation allows only 5.0 g, dry it again and reweigh before you report anything.
Think of it like this. Actual yield is the juice you really pour from a bag of oranges, compared with all the juice the oranges contain: some stays in the pulp and some drips onto the counter. You only learn the amount by measuring the glass. If the glass holds more than the oranges contained, something else got in.
Core idea. Actual yield is measured, not calculated: the mass of product really collected after a reaction. It sits below the theoretical maximum because of losses, and a value above the maximum signals a wet or impure product.