I've been doing a lot more "longer" races on Zwift over the last 2-3 months to get ready for the Chasing Classics.
One thing I've noticed is that I might be able to hang with the fast guys for longer than other people in my "Cat" but many of them have learnt how to pace sustainably over that duration and catch me later on when I'm flagging.
Durability also explains why training on these longer 1.5-2 hour races makes the shorter DRS 45 minute races "seem easier".
I've always found FTP tests overestimate my true capability. Like CP and W', it's all modelling of complex biological systems and we all have different physiology, nutrition, stress, health, sleep, fatigue, and cumulative training time in the legs etc. so there will be errors.
I think Eddie Merckx had it right. Ride your bike a lot.