A particle in a box exerts pressure on the walls of the box. This makes sense from a classical perspective. A particle that's bouncing back and forth in a box will exert an average force on the walls over time. But it also makes sense if we model the particle as a harmonic oscillator. If we increase the width of the box, the energy of the particle inside will decrease. The fact that there's a change in energy tied with a change in distance/position means there's a force at work. And it turns out that the bouncing particle model and the harmonic oscillator model give the exact same results mathematically.
The reason I think this might be similar to virtual particles is that the harmonic oscillator model represents what's happening with the particle more accurately, but the bouncing particle model might be easier to understand. This seems to be the case with virtual particles. Real particles aren't exchanging virtual particles one by one to exert force on each other as Feynman diagrams would suggest. But there is a change in energy associated with a change in distance, so there must be a force involved. But that can be really hard to conceptualize, so instead we act like they're exchanging individual particles so we can write simple equations and make simple Feynman diagrams.
I'm not suggesting that this is a perfect parallel. But is it on the right track?