Someone did a research study on this
(would you believe) and discovered that collisions were most likely when there were two drivers in very similar cars who were themselves of very similar performance. The analysis went further
, indicating this is applicable to workplaces in general, and that one should watch out for one's "work twin" if avoiding overcompetitiveness. I strongly recommend reading the analysis at least, as there's lots of interest for the racing fan.
If new drivers are going to teams where their new team-mate is significantly better than them, then all is sunshine and roses. Unless there's a similarly-performing car with an equally good/bad rookie.
If new drivers go to teams with said rookie as their team-mate, or they find the experienced team-mate isn't really any better than them, or they both discover the car isn't worthy of them (thus forcing them to the same level at the car's "ceiling"), things tend to go pear-shaped for collisions. This tends to result in either demoralisation (potentially causing crashes through inattention) or anger (this is when they tend to hit each other through either the ignition of competitive streaks, or as a substitute for lobbing the car into the nearest skip).
Now, we can't yet know which cars will be disappointments (and recall disappointment is the gap between expectation and reality, so it's easier for Mercedes or Ferrari to be a disappointment than McLaren or Williams). What we can guess is which partnerships are likely to cause trouble:
No disrespect to Valterri Bottas (especially as I recently found out Dad's a fan of him), but he's not currently in the same league as Lewis Hamilton. Because of this, team-mate collisions are pretty unlikely.
Well, these two should be far enough from each other for the described effect not to kick in, but with Leclerc hitting the Sauber's performance ceiling more often than not, it's hard to tell. Maybe
Red Bull: Verstappen-Gasly
I think these two are closer in skill than either of them thinks at the moment. Combined with the obvious favouritism from Helmut Marko, and this is a recipe for fireworks. Yes.
Toro Rosso - Kyvat-Albon
This is a difficult one to read, because both have faced substantial problems on the way up, and I'm not convinced either is proofed against the "tricks" in the Red Bull way of doing business. The team atmosphere is enough ceiling for either of them to hit. Yes.
I don't worry about these two against each other. Not because they're very different in performance but because they tend to perform at different races. They're hardly ever both "on" or "off" at the same time, which is necessary for collisions to be a threat. No
McLaren - Sainz-Norris
I think Sainz will be a bit better than Norris, and both are expecting McLaren to be at least as weak as it is likely to be, if not weaker (therefore, they may be pleasantly surprised, and certainly not disappointed). - No
Force India - Perez-Stroll
While I'm hearing signs Stroll is improving now that he's away from Williams, I still think Perez will be stronger. Stroll isn't Ocon and getting that level, if it's even possible for him, will take quite a long time. - No
Williams - Kubica-Russell
I think Kubica is significantly better than Russell, therefore driver strength isn't similar enough for this to be an issue - No.
Renault - Hulkenberg-Ricciardo
Now this is a line-up that will bring fire that wasn't there before. I believe Nico is capable of more than he's shown, and that someone like Daniel is exactly who can cause him to extract it (because Nico's previous best came when he was paired with an equally-strong Sergio). However, if he does make that step, he's going to be in the neighbourhood of Daniel's performance level, and therein lies a problem... - Yes
Sauber - Raikkonen-Giovanazzi
This is the lineup with what I believe will be the biggest disparity. As such, it has the least chance of collision.