Both of these numbers very close to what one would expect.
The number of combinations possible for 5 numbers in a field of 55 numbers is 3,478,761. The number of possible combinations of 5 possible numbers in a field of 31 is 169,911. Thus the subset of combinations covered by the 31 field is 4.9%, close to what is found.
The number of possible combinations covered by a field of 39 numbers is 575,757. Thus the percentage of this field's combinations covered is 29.5%, close to the experimentally determined (i.e. drawn) value.
People may imagine that by "spreading out" the field, it is more probable that they will win the lottery, but this is just wishful thinking. In the attempt to "spread out" they are, in fact, placing similar restrictions on the probability of winning and such restrictions will be just the same as the restriction of choosing numbers that are 31 or less. It is just not as obvious that this is the first case, but it is the true all the same.
There is no way to increase one's odds of winning the lottery, although there are several ways that one can decrease one's odds per ticket, if one is purchaing more than one ticket.
If one purchases two tickets, for instance, one would expect that one is cutting one's odds by 50%, but this is not always the case. Sometimes you can buy two tickets and decrease your odds by less than 50%. Suppose that one bought two tickets in a lottery for which the odds were 1 in 100,000,000 but that the two tickets were identical - the purchaser had chosen exactly the same numbers. The odds of winning would have not decreased at all. They would still be 1 in 100,000,000, although should those odds be beaten, the owner would be entitled to two shares of the prize.
It seems to me that this is what people are missing when they "wheel," at least as I understand wheeling. Wheeling it seems to me is a way to guarantee that one will not decrease one's odds by the factor equal to the number of tickets purchased.