Experiment to Prove/Disprove Rubber Banding

There are two threads and in both the existence of rubber banding is being debated. I think we can prove or disprove it’s existence and it requires a set of tests being executed by a group of players of various skills. If two players of different skills use the same stock car on the same tracks for the same weather conditions, the pace of the AI will vary if there is rubber banding. The AI will be faster to catch up to the faster player.

So, I’ve created six Blueprints with a Circuit and non-Circuit route on the Road, Dirt and Cross Country all using the 1958 Morris Minor 1000. I’d propose some different drivers of different skill levels use the same stock car with no tuning on those circuits on Average difficulty. Assists don’t matter. The reason I chose a slow car is in Forza Horizon 3 when I experienced rubber banding it was most evident on a D100 car. Plus in a slower car, a faster driver will take longer to get far enough ahead to lessen the opportunity for the AI to get bored and stop trying as I’ve seen in previous Forza titles.

If you guys all think that using a higher or lower difficulty will expose rubber banding I’m all for it. However, if rubberbanding exists we should see substantially different times from the drivatars for the top three cars.

The events I created blueprints are:
Horizon Festival Circuit
Horizon Festival Spring
Ambleside Scramble
Derwentwater Trail
Gardens Cross Country Circuit
Glen Rannoch Cross Country

All I need from everyone is the times for the top three cars. If you finished outside the top three. Give me the top three and your time. If you are in the top three, just give me the top three and tell me which of the top three positions you were in. For circuits, I’d like to see the Best Lap and the Time, but for the other non-circuit events, just the time.

This is the only way I can think of to gather evidence for or against the existence of rubberbanding in Forza Horizon 4


You really only need the one point-to-point track ‘Horizon Festival Sprint’. I was able to experience rubber-banding there (Finally) I ran about 10 tests just to knock out any variation.

2 tests where I win by 3+ seconds
2 tests where I hang in the middle of the pack
1 test where I wait 30 seconds.

Then I switch to a high S1 vehicle to test that.
2 tests where I win by 7+ seconds
2 tests where I hang in the middle of the pack
1 test where I wait 30 seconds.

HOWEVER, forums have complained about it being ridiculous and unbeatable, when I feel it’s quite the opposite. If you have no problem getting in first place, you should have no problem finishing in that position. During the Morris D100 races, I would pass first place and that was it, It did not appear as if they were gaining an unfair amount of speed to catch me by any means.

1 Like

Your findings match my own. Unlike FH3, once I’m in the lead, if I maintain my pace, I’ll win the race. I’m actually finding it’s easier to get through the middle of the pack in this version, as I’ve said in another comment, I think because the cornering of the AI cars has improved, selection of the right car and/or tune is key with the higher levels of difficulty because the AI isn’t using spikes of HP to catch up if AI falls behind. The AI can drive faster and smoother in the corners, so assuming you have a good match with PI, you have to drive faster with better turns.


FH4 uses pretty much the same AI system as Grid Autosport whereby no matter how fast or slow you drive they will always be there or thereabouts to your position on the track. If you have an accident they will pass you and then slow down according to your difficulty setting. That’s an adaptive AI system. The only flaw in this system (in FH4) being that if you miss a checkpoint towards the vey end of the race you will never catch up again.

Rubber banding, as seen in TC1 and 2, NFS and Burnout Paradise is when you get a certain distance in front or behind (the stretch point of the elastic so to speak) and in the AI react by catapulting themselves towards you in a very short space of time or unnaturally slow down and allow you to catch up in an equally as short space of time. In my 300+ races in FH4 I have yet to see this happen on the difficulty setting I am using. Yes, they sometimes give the impression that they’ve caught up unfairly (usually when you have a slightly slower lap) but that’s purely down to the fact they were never really that far behind in the first place.

(PS) Try the Horizon Festival Sprint in a Unimog on ‘unbeatable’…A huge “oops” on the part of the devs but highly entertaining to watch :slight_smile:

I think what some of the users are experiencing is flaws in the calculation of the PI for AI cars. So, some cars the player might pick or build or get a tune are over matched by the PI selected for the AI competitors in the race.

For example, in Cross Country races I was struggling to beat Average difficulty, but with the blueprints I setup for this test I was easily beating the AI in higher difficulty levels. The AI was restricted to the same car as me. I haven’t tried yet, but I have a feeling I could win in Unbeatable mode with those blue prints using the stock Morris Minor, however, I know that with A or higher tuned cars, if I didn’t pick the right car, I’d struggle to win on Unbeatable because the AI’s PI would be better suited to the race and now that the corner speed differential with the AI is not as great, my mistakes in driving and car/tune selection have a bigger impact.