It's more reliable to "torque a bolt up" using angle of turn than torque. Also reduces the effects of friction of the model. Friction tends to pretty uncontrolled in the in field- a thread could have dirt, swarf, and/or varying amounts of oil.
consider a nut running on a bolt, with a compression spring between washers and the nut and bolt head. If it's a very weak spring very little tension force in the bolt shaft will be needed to balance the compressive force, so you can consider the bolt to be stretched by a negligible amount . For each complete turn of the nut relative to the bolt head, the nut moves closer the bolt head by the screw's *lead* and that compresses the spring (for single start thread, pitch=lead) . You can use Hooke's law (F=kx where k is the spring stiffness and x is the length of compression) to work out the compressive force on the spring and the tensile force acting through the bolt.
Now replace the bolt with a tension spring, fixed between the washers. Assuming all stretch, occurs between bolt head and nut, that equivalent spring has uncompressed length equal to *about* the distance between bolt head and the nut. it also has the same stiffness as the bolt. You have to solve F=k_1*x - k_2*x where K_1 is the tensile stiffness of the bolt, k_2 the compressive stiffness of the clamped material, and x is the distance the nut has to have moved along the *stretched* bolt for a given angle of turn.
Ok now onto errors. I've already mentioned not all stretch occurs between the bolt head and nut. some will happen beyond that distance. and under very heavy forces the parts might even "bow" a little. with high stiffness they don't need to bow much to make a huge difference in applied force.
some slack due to surface roughness between washer/nut/ and bolted surfaces might be squashed out.
When there's lots of friction acting on the threads, not all the "turning effort" will make it through to the bolt's shaft- the bolt will be ever so slightly twisted- again that changes the stretched length. Same can apply if the bolt head binds, and the nut binds to the bolt for a short time- that can add a twist to the bolt shaft.
temperature- the bolt and the material to clamped are likely to be of different materials. unless they are torqued up at service temperature, the torque will change as the materials expand at different rates. If a bolt is torqued up at room temperature,and it expands less than the bolted material when heated up to higher temperature, that exerts more force on the bolt shaft- if that force gets too high the bolt will either snap, or more likely it'll exceed Hook's limit, and have a permanent extension of length. When cooled back down that permanent extension of length will have reduced the torque- Suspect that's what was going on with the situation described by jimmymae2000- there's also situations were vibration as likely in engines can also knock out some of the "twists"/settle the parts into position, might also turn the bolt heads/nuts if they ain't locked in place using lock wire or similar