[POLL] Are you able to reliably select an AI level that matches your ability?

[Shovas]

Are you able to reliably select an AI level that matches your ability for the typical 80-120 or 0-100% selectors? Please comment whichever poll option you choose. I really want to know how this isn't driving people insane.

The ability to reliably select AI level is something that is driving me up the wall - in fact I just can't do it. I'd love to know if and how others are successful at this. I find picking an AI level only at the beginning is just a shot in the dark and you end up improving your laptimes as you practice and qualify anyway and then you need to start the whole session over again and pick a new AI level. This is especially problematic for championships, say ACC, where an AI level applies throughout the whole thing and you end up with good AI at one track and incredibly slow at another and then way too fast on yet another.

This is a big problem for sims that aren't iRacing where the multiplayer infrastructure isn't there. And even if it is, and even for iRacing, AI racing is still important because of lack of frequent race sessions that are available to you throughout the day.

Evidence Based AI

I also want to recommend what I'll call "evidence based AI" which boils down to checking your laptimes in practice and qualifying and using that as an average or median to calculate the AI level from. When training AI they're running laps anyway why not record the laptime so that it can be compared against the players? I mean they probably already do. I'd love to hear thoughts on this.

The idea above is similar to the manual "spread" you can do in AC Content Manager where you can select a lower and upper bound and it will spread out the AI within those levels. The problem is it's still manual.

 

[Clover11]

This might be an impossible target to hit for developers, but I like your idea of varying it based off your practice and qualifying times. Part of the challenge is that I am always going to be better at some tracks then others. Plus, what I want from the Ai may not be what you want. For me, I like to win but I want to battle my way to the front. Ideally, I mange to make my way to the lead with a full lap left for me to hold off the pack. Some people may want to lead for more and some want a tough a battle as possible without the need to win. Maybe they could tie some of it together so that if you get faster, the compaction improves by a similar margin.

 

[Shovas]

This might be an impossible target to hit for developers, but I like your idea of varying it based off your practice and qualifying times. Part of the challenge is that I am always going to be better at some tracks then others. Plus, what I want from the Ai may not be what you want. For me, I like to win but I want to battle my way to the front. Ideally, I mange to make my way to the lead with a full lap left for me to hold off the pack. Some people may want to lead for more and some want a tough a battle as possible without the need to win. Maybe they could tie some of it together so that if you get faster, the compaction improves by a similar margin.

I like your ideas. I've thought about those cases, too. I think you could easily provide a couple more selectors for a rough position on where you'd like to grid in the pact: Front, Middle, Rear. Or you could phrase it like "Chance to win", "In the heat of it", "Backmarker".

The rFactor 2 AI Limiter option also helps in that it spreads out the field or keeps them closer together which can help you feel like you're having a more realistic race depending on what you like.

Lots of very simple ideas could be done and all infinitely better than the overly simplistic single AI level option we get now. It kills me. I just cannot find the right AI level and I keep restarting and adjusting the level to try again but it's still off because I keep improving. It's crazy.