With so many options for calculating the rate, it is important to know when to use each method.
Regular rhythms
If the rhythm is regular and not too fast, you can use any method to calculate the rate. Counting the large grid squares is probably the most commonly used method.
Irregular rhythms
If the rhythm is very irregular and there is no clear pattern, you should calculate an average rate. You could also describe a rate range by using the two closest and two farthest apart beats, e.g. Atrial Fibrillation could have a rate of 120-160 bpm.
Very fast rhythms
If the rhythm is very fast, the large squares may not be very accurate, so use the small squares or an average rate.
Other methods
There are various other ways of finding the rate that are derived from these methods:
ECG rulers
ECG rulers are printed with a scale that is calibrated to the standard grid squares. If you line up a beat with the start of the ruler, you can read off the rate at the next beat. This is the same as calculating a beat-to-beat rate using the large squares method.
Memorise the pattern
Once you've done a few examples of calculating the rate using the large squares method, you might start to memorise the sequence 300, 150, 100, 75, 60, 50... You can quickly approximate the rate by counting this sequence out across the large squares between two beats. This is the same as using the large squares method.
Can I just trust the computer's calculation for the rate?
Computer interpretation of ECGs is notoriously unreliable. Even though the computed rate is not as unreliable as some of the other ECG features, it can still be tricked by artefact or very abnormal waves such as tall T waves. If you are going to use the computer interpretation for the rate, it is worthwhile knowing how to quickly estimate the rate (e.g using the large grid squares or average rate method) so you can check the computer is in the right ballpark.