If you take 2 groups of equal cardinality, it could happen that, for example, many of the higher values go to the first group and therefore, many of the low values go to the second group, making their respective means quite different and different from the original sample mean. This could even go worse due to the possible existence of outliers, that is, values that are far different than the sample mean. An outlier tend to disrup the mean of a sample, but for smaller samples the result is much dramatic.
For example, let X be {1,2,3,4,5,6,7,8,9,100}
The elements of X sum 145, hence the mean of X is 14,5. Let divide X in two groups
Y = {1,2,3,5,9}
Z = {4,6,7,8,100}
The elements of Y sum 20, so its mean is 4
The elements of Z sum 125, so its mean is 25
Both means are quite different from each other and quite different from the mean of X. Note that if we take the mean of the means the result is 4+25/2 = 14,5 which is equal to the mean of X.
×
= 
