Okay, so there are times when products are added to cart accidentally and
then need to be removed. Undesirable. And if you can try to
ameliorate those occurrences great. But don’t break something else while
doing so.
If you only look at the 1 in 10000 times this mistake happens and say
“Well, 100% of the time that the user adds then removes an item they
performed the wrong action so let’s fix it” is looking at the data in a
vacuum. The correct analysis is 9999/10000 times the correct action
happened. If we can fix the 1/10000 without regressing the 9999/10000,
great! But you’ve taken a 1/10000 problem and now you have a much higher
percentage-of-the-time problem instead! This is a significant and amateur user
interface and data interpretation fail.
If you don’t like percentages, the same analysis can be performed on a
time-wasted basis and it leads to the same conclusion.