We know that outdoor transmission does happen, but at significantly lower numbers than indoors. In fact, the initial outbreak at the outdoor Wuhan market is now considered to be the first identified super spreader event rather than where the virus jumped from an animal host to humans (which happened earlier). It's important to understand the circumstances where outdoors transmissions have occurred, and have measures to reduce those circumstances as well as communicate safe behavior to potential participants in a race.
A local-to-me spreader event was an outdoor party at someone's house, where people weren't wearing masks. I think it's likely the virus was spread due to conversations - prolonged talking that puts droplets/aerosols into the air into each others' faces. It's the same reason bars been hotspots for transmission indoors (though bars are much risker because of the indoor factor). Maybe they would have been fine if it had been windy that day.
What you want to avoid at a race (or training group) would be those close quarter conversations with other people while unmasked. I'll throw in heavy breathing, since that hasn't been studied that I've seen. I watched the Seth James Demoor video of the Kendall Mountain Run, and the measures that the organizer took seemed reasonable to limit that transmission route: Masks wear required at the start and finish, when people would be having conversations in potentially still air, and clear plastic hanging down between participants and people manning the pickup table. Also, having staggered 10 person waves limits the number of people together having conversations.
During the race itself, the 10 person waves spread people out, and you have apparent wind caused by running itself, so the possibility of transmission should be minimal. However, I wouldn't bet that you'd be safe if you were in the slipstream of an active spreader for the whole race. So, if I were race director, I'd maybe say no drafting (unless everyone has been tested, like the BTC time trials), run alongside someone if you are going the same pace.
Thinking about a mass start, I'd think the main risk would be the crowding at the start, and some people probably tend to follow a specific person's slipstream for a prolonged period. Though I can't quantify how much, it's obviously more risky than a small wave staggered start. This is using the same reasoning behind limiting and scaling sizes of gatherings depending on how bad it is in a particular community.
I personally haven't worn a mask outdoors running or cycling. It's not required outdoors where I live, and briefly passing someone outdoors is nothing like having a conversation. The case studies and guidelines for contact tracing all have an element of spending a decent amount of time, unless coughed in the face or something. The CDC says 15 minutes within 6 ft. I don't think that's conservative enough, and contact tracing in some other countries is tighter than that, but passing someone briefly on the trail outside should be safe.