Introduction: Scotland 3 is a very slippery version of Scotland 1. The track has more obstacles and the Stormy weather makes for some interesting AI behaviour. The results in the bottom half of the field are certainly rather unpredictable, whilst the victory is highly contested by 2 AI in particular. I will delve into each of the eight different AI drivers' performances in Scotland 3, including anything noteworthy. A reminder to all that my findings apply to Total Drivin and AI performance can vary slightly between Total Drivin and Grand Tour Racing '98. For a new player this blog will be difficult to understand, and I would suggest visiting this page to comprehend it better:

AI Xu: AI Xu has dropped the ball somewhat since his last outing in Scotland 2. The slippery conditions seem to be an issue for his slightly inaccurate driving style, sometimes showing good results but otherwise performing worse than expected. He and AI Ivanov are very similar in pace, although he usually finishes ahead simply because he starts higher on the grid.

AI RossiHere we have one of the two contending AI in this race who fight for victories; AI Rossi. He is back to winning form after a slight decrease in ability in Scotland 2, with an exceptional charge on Lap 3. He fares very well in the wet conditions, often making up a lot of ground early on as the drivers with lesser accuracy perish. Like in Scotland 1, his ability to finish ahead of the top AI is dependent on how quickly he can pass the slower drivers.

AI Morgen: AI Morgen once again fails to perform. His erratic driving style combined with speed which he lacks control over sees him constantly colliding with walls and being overtaken by more accurate yet slower drivers (i.e. AI Ivanov). He is probably the worst AI driver in this level.

AI Ahmed: Scotland 3 is a breakthrough level for AI Ahmed. This is the first level where we really see that he is a much more capable driver in Scotland than AI Lumiere. His very accurate driving style like AI Rossi means he is able to overtake the French driver quite easily, only failing to on the odd occasion. 

AI Baptiste: AI Baptiste's form varies significantly in this level. If he can maintain good car control then he can finish in the middle-upper positions. However, as soon as he starts to slip up due to inaccuracy he will be overtaken by drivers such as AI Xu and Ivanov who commit less mistakes. 

AI IvanovIvanov's AI is much more impressive in this level than in Scotland 1 and 2. His driving accuracy seems almost on par with AI Xu, and together they are able to capitalise on errors from Baptiste and Morgen. However, starting as 5th HT (normally) reduces his speed signifcantly in comparison with drivers like AI Xu and this limits his progress. Nonetheless, it's a clear improvement. Ivanov's Out in SC3-4 (shown in the below table) was an unusual anomaly with no clear explanation as to what happened.

AI Roberts: Once again we find AI Roberts at the front of the pack. He has a lightning speed for the first 2 laps, completely unthreatened by the more sluggish AI Lumiere. However, in Lap 3 he slows down a lot, in a similar fashion to Scotland 1, which allows AI Rossi to catch up. There appears to be a 50/50 split between an AI Roberts victory and an AI Rossi victory, both exceptional in this level. 

AI LumiereAI Lumiere's performance in the wet conditions is poorer than AI Ahmed's. He is highly likely to be passed by the Egyptian driver during the race and will finish behind both him and AI Roberts. If the player were to choose Ahmed then AI Baptiste would not pose a serious threat to the French driver as 3rd HT in the place of AI Ahmed. 

Here is my table of results which I created during my study of the AI Behaviour in Scotland 3. I chose a different team each time, allowing the AI to race without interference from myself and recording where each AI driver finished. SC3-1 is the first race and SC3-2 is a new race having restarted with a different team and so on. N/A indicates the team I was using for that race. 

Scotland 3 AI Results (by Jbracer2)

Community content is available under CC-BY-SA unless otherwise noted.