It is impossible to judge potency (effectiveness per milligram) by that method, saying that one steroid is like another but for one methyl being added, or that sort of thing.
There is actually a quite dramatic change in the shape of the molecule that results from adding a 17-methyl. The entire structure twists in a complex way to accomodate it. Receptor binding is a function mostly of shape. The human mind cannot predict in detail either the exact shape change or, if it knew the exact shape change, the resulting change in receptor binding.
For example, could you predict that the oxandrolone structure, on losing the 17-methyl, would lose absolutely all potency? Even when taken by injection. I don't believe any human mind could predict that.
So that is an example of a molecule going from zero (with no 17-methyl) to fairly potent, "simply" on addition of this methyl.
There are other examples of a 17-methyl adding no potency at all where the comparison is by injection in both cases.
It's not predictable.
The same is true for any unusual per-milligram toxicity. A good example is methyltrienolone. One couldn't deduce from the structure that it would be tens or hundreds of times more toxic per milligram (if it is; it is claimed to be though) than pharmaceutical oral anabolic structures.
So, as another example, reasoning that "Well, Dianabol is essentially boldenone with an added 17-methyl and can, for reasonable periods of time, safely be taken at similar total dosage per week as is typically used for boldenone, so surely adding a 17-methyl to trenbolone would yield a compound that could be taken at similar total dose per week as trenbolone" would be quite in error.