MotoGP 21: Milestone needs to address AI pace balance

MotoGP 21 came out last month and we love Milestone's latest official MotoGP game. To find out exactly why, check out our full review of the title.

While we gave MotoGP 21 a positive review, the game isn't perfect and there are some issues that need to be addressed. One is the AI, which despite much improvement, are still lacking compared to other sim racing titles.

So, what's wrong with the AI? And what can be done to rectify it? We've got everything you need to know right here!

The problem

The heart of the problems with the AI in MotoGP 21 is that difficulty settings dictate their lap times. That sounds obvious, but there are several ways of going about achieving this. The timing screens won't reveal any issues, but playing the game and spending time racing against the AI will unearth some problems.

MotoGP™21 20210429092133
expand image
PROBLEMS IN THE TURNS: The AI are fine on the straights, but it's the corners where they struggle

If you set the AI to their highest difficulty, or near the top, you won't notice too many abnormalities. However, for those that are on a lower difficulty (around 60 or lower) certain problems become evident.

In order to scrub speed off the AI, they've been slowed down the straights but not so much in the corners. It's not the case that no speed is lost in the corners, it is, but the balance is off. If you turn the AI down to near or at 0, the problems become worse.

You race past the AI on the straights with ease, but when you get to the first corner, the AI are still almost as quick as a skilled MotoGP rider. Now, this may not sound sinister, but it does have knock-on implications for racing.

Bonkers behaviour

Having AI off the pace in areas of the track really makes you re-evaluate your approach. Anybody who drives a car and gets caught behind someone doing 20 mph in a 30 zone will know what we're talking about.

Being slow in the corners ironically makes overtaking and racing with the AI more difficult, especially when they're in a group.

MotoGP start jerez
expand image
THROUGH THE FOG: In the rain, the problem becomes worse, as visibility is lower

The biggest problem though, is when the AI try and overtake you. Because they're so fast in the corners, the AI routinely divebomb the player, usually ending with a crash and both riders falling off their bikes. Sometimes the AI get overzealous and dive down the inside so fast that they spin by themselves.

This is still a problem on higher levels of AI difficulty too, it usually occurs when the player is close to the drivers around them.

All of this ruins the immersion of what is overall an excellent game. It takes a lot of effort and work to create an experience that feels real, but only one aspect has to be off for the whole house of cards to fall.

Remember Neural AI?

A feature of MotoGP 20 that was promoted heavily was Neural AI. However, we hadn't heard anything from Milestone regarding the feature in the leadup to MotoGP 21. Could this be the source of all the AI's problems?

It's certainly possible, as the whole point of Neural AI was that it learned from your driving style and adapted itself to counter it. Perhaps Neural AI 2.0's findings were collected by Milestone and used to adapt the AI in MotoGP 21. Going off what we've seen from MotoGP 21 though, that's questionable.

How to fix them

It should be a relatively simple fix for Milestone. Increasing the AI's speed down the straights and decreasing their cornering speed should allow them to feel more authentic. It's not as straightforward as changing a few lines of code, but for an experienced company like Milestone, it's very doable.

In any case, Milestone can fix these issues with its AI by patching MotoGP 21. However, with Hot Wheels: Unleashed on the way later this year and MotoGP 22 already likely in development, it's doubtful as to whether the Italian developers will have the time to do so.

This Article's Topics

Explore new topics and discover content that's right for you!

MotoGP
Have an opinion on this article? We'd love to hear it!