I've been designing wargames for a while now. Back in my salad days - during which time I was actually a lot more carnivorous than I am today in my non-salad days - I was a little obsessed with obtaining an accurate order of battle, and more than a little terrified that I was going to botch it. I assumed that I would make some seemingly small mistake - a misplaced regiment here, an imagined brigade there - and that my design's ability to function as a simulation would be fatally compromised.
This panic was only helped along by the fact that in my first few games, the source materials were contradictory as to who was where and when, and it seemed like no matter what decision I made, it would be wrong - that there was no way to arrive at the definitive truth, or if there was, I wasn't smart enough to suss it out. I'll spare you a detailed mapping of my neuroses, and skip ahead to the part where I made my best guess and crossed my fingers, hoping that either (a) I was correct, or (b) if I was wrong, then no one would find me out.
For my first game, the answer was either (a) or (b) - certainly hoping it was (a)! But in one of my other early designs, I made a pretty embarrassing order of battle error, when I had one of the units begin the game a couple hundred miles from its actual starting position. And, sure, that one unit wasn't anywhere near the frontline, didn't see action historically until a few turns in, and would be able to get to its position long before the enemy broke through the first line - but it was still fairly mortifying!
But, you know what? That one misplaced unit didn't really ruin the game. In fact, I had it out of the box a few months ago and the thing still zips right along, still trends toward the historical result, and still replicates the historical factors. That's because the game's success or failure as a model and as a game didn't really depend on the fidelity of any one counter or its combat factor. A small error in the details won't substantially alter the result - it doesn't matter if you make some nettling little mistake. This realization really helped me ease up on the panic attacks and helped to alleviate what the kids these days are calling "impostor syndrome".
That's not to say that the order of battle itself doesn't matter, or that one can take a cavalier attitude to research. If there is a substantial error, or a substantial number of them, then it will throw things off. But an orbat by itself isn't a game design, and a good game design doesn't rise or fall on the strength of its orbat. It's worth spending the time to try and get it right, but it's not worth giving yourself an aneurysm in the process.