This is something that I have a bit of difficulty with. How does everyone select the mons that (insignificant) enemy trainers use? I'm not referring to obvious factors such as level curve or base stats. What I'm asking is how do you choose whether that youngster has two Zigzagoons, two Lillipups, or perhaps one of each? When you enter a forest, how do you decide which bug catcher has Scatterbugs, which has Wurmples, and how those trainers are organized in terms of the overworld? Do you try to make specific trainer classes have a select few evolutionary lines, or do you make most trainers have mons based on nearby routes and caves? At the moment, I don't have very specific guidelines. All I have to go on is what "feels wrong" when I test. If anyone has a better methodology, I could use some help.
For me, it's a mix of things. Early on, there are basically the criteria that I weigh it on: what's in this area, what trainer type are they, and what will the player have up to that point. I'd imagine that the third point is often overlooked or, in cases of "uber hard" hacks, abused, but by weighing all of these things you get a pretty clear idea of who should have what, when. Other things to take note of is how long you want the route to be and how many Pokemon do the other trainers have, as several teams of the same or similar Pokemon can be incredibly monotonous.
As the game progresses, you can throw out focusing on what the player may have (keep the starters in mind, though, as for most players this is a constant). From here, the trainer's type and the area type become a lot more important and what you should be focusing on is difficulty variance. If you have a cave with five Hikers, don't make them all the same. Their difficulty should be balanced with their position. The way the official games do it, the harder trainers are generally the more optional/avoidable ones or near/at the exits to their locations, while the easier ones are mandatory or close to the entrance. This is a pretty solid way to do it as it gives the player time to prepare without necessarily allowing them to bob and weave through the difficult enemies easily. Generally, it takes thought and timing.
As for what they should have in those instances, it can be a combination of things. Though not as much as before, what's in the area is still relevant and you should delve into that. These Pokemon may be themed, and because the player will likely be prepared for these, they should be issued mostly to the lower difficulty trainers in said area. This is good, because since these guys are generally mandatory or near the entrance, it gives the player a taste of what they can expect from the wildlife and other trainers (which can be a good motivator). If you have trouble with picking Pokemon, try starting by looking at who the trainer is, where they're positioned, and what you as the designer want out of them. If you want a trainer that's a moderate challenge, why not start them off with a pokemon from that area and then build around that?
Themes can also be good, and the official games have fun with this. Sometimes, for specific or random trainers, the games will give teams specific themes (all the members of an evolutionary family, Zangoose and Seviper, all of the starters from that region, stuff like that). They're a nice fallback if you're out of ideas and the player will get some appreciation out of them, but they're best used sparsely (assuming they aren't relevant to what's in the area) because the novelty will definitely wear off if used too often.
And for the endgame the important things to note are trainer type and teams that compliment each other. Here team making must be more efficient because, by this point, moves pack much more of a punch and the player will be using final evolutions and legendaries. This isn't to say that every match should be a challenge- no, the curve should still exist for difficulty, but you need to make teams that are competent enough not to simply be steamrolled by the endgame pacing that the player will bring with them.
