You might want do a search on the topic. It has been beat to death. No one has ever made a definitive case as to what frame frame connectors actually do IMO. EVERYONE has a opinion. There are many forces acting in concert here. Twist/chassis torque certainty is one, and frame connectors IMO really don't do a good job of resisting the twist, help, but a poor solution, and certainly better then nothing. Twist would be best resisted by a 5'? diameter tube (obviously impractical), and 2 2x2 or 2x3 tubes on one face of a rectangular box are not an efficient design to resist chassis twist. But to answer your question where do forces go, if the chassis does not absorb them, they remain useful in accelerating the chassis. The chassis will release only some of the stored twist if its beyond elastic state, but unless its a 100% efficient spring, (its not) some energy will be converted to heat as the chassis bends/twist, and will not given back as the power comes down, but this is likley splitting hairs.

Reality check, that half the population is smarter then 50% of the people and it's a constantly contested fact.