This is a major difficulty with standing armbars. It's not that they can't work, because they certainly can. The problem is that because you have less control of your opponent, they only work against a competent, fully resisting opponent if you apply them full speed, which doesn't leave time to tap. That means you can't really test and polish them properly in sparring (assuming you don't have an unlimited supply of training partners willing to have their arms broken). If you are getting taps with a standing armbar in sparring, it means your partners are tapping way early and not fully testing whether they can escape the lock. A lot of the skill involved in reliably finishing a lock comes from trial and error, finding all the subtle points of failure in the last 10% of the technique where you thought you had your opponent caught but the angle or the position or the leverage weren't quite right. With a partner tapping early to your standing armbar, you never get a chance to develop that skill.
This is one reason why, out of thousands of MMA fights, I know of only one that was finished by standing armbar (Shinya Aoki broke Keith Wisniewski's arm with a standing Waki Gatame in 2005) compared to hundreds of finishes via Juji Gatame on the ground.
(The other reason is that the increased control involved in ground submissions makes them inherently higher percentage even if skill levels were equalized.)