Nonono… That’s absolutely wrong. The odds of ten independent draws of 1% is most certainly NOT 10%!!! It’s lower.
This is compensated by the multiple levels you might go up, if that wasn’t capped at 7!
Example to clarify:
A. 1 draw 100% =1 level up
B. 2 draws 50%= 0 level 25%, 2 levels 25%, 1 level 50% => On average the same 1 level.
But if we would do this excercise 7 times simply to get to the needed 8, method A. would guarantee the end goal, whereas method B most certainly wouldn’t be a guaranteed succes. Indeed, there’s even a very remote chance method B yields no levels at all. OTOH, even with the wildest luck in the universe, you would not get 14 level ups… 7 is the best result you would be able to get. IOW, method A would be by far the better option.
Now this is a very simplified example using 50% and 100% draws, but it shows the mechanism and why the ten draw method is superior. Considering we need to get 7 level ups, maximising the odds in a draw is better than gambling on multiple level ups with lower odds.
The reason why we still stick with 1 star heroes is the scaling of exp would limit the draws if we use higher feeder heroes (We reach the xp finish line too soon, which affects the efficiency of the excercise). That said, if it was just a matter of dinging special skills with no other consideration, maximizing the odds is clearly the way to go.