Board Thread:Game Discussion/@comment-35745775-20180602191223/@comment-35745775-20180603100906

What has been said already is all correct, and yes I am familiar with the TSM algorithm, the bot list here in the Wiki (thanks again Amrosa) and understand how winning and losing (or winning with slower times) affects the bots, but what I have found out I have not seen anywhere yet.

It is a bit complicated to explain, and as Amrosa said, basically I am beginning to reverse engineer some of the game mechanics.

So for those who either, a) have an interest in finding a way to measure ability (complimentary to WTTT groupings) objectively. That is incidently what the game algorithms are trying to be all the time. b) would be interested in monitoring their own progress or, c) have an interest in better understanding and investigating the game mechanics, then I think it is worth while digging into more.

I have only investigated the Nissan Skyline and BMW M3 ES so far, but the parallels are enormous. This also all only applies when you race online.

Starting with cup races in the Skyline ES: With a 22 car grid, you will always find 9 offline bots (plus 12 online bots, plus yourself). The top offline bot is most likely the current reference bot for the series, though there are some exceptions. The offline bots are always 2 bots apart from the bot ranking list, as found in this Wiki.

So an example: Tier 11.1 in Skyline ES, Cup race @ Catalunya. Let‘s say your current reference bot is Eraser. The grid line up will include the following offline bots: Eraser (0 = ref), Richie Rich (-2), fortz99 (-4), JimBling (-6), _thefreak_ (-8), Bronson Gulthry (-10), PerturB (-12), morkRING (-14), Oddly (-16).

The grid positions and finishing positions of those bots is totally irrelevant for this exercise.

All of the ‚standard‘ cup races in this series will always contain those bots, so long as you don‘t race and influence your reference bot. Just try opening Tier 9.1 (Daytona Motorcycle) or Tier 8.3 (Suzuka East).

Then come the ‚short‘ cup races in this series: Tier 1.2 (Dubai Grand Prix), Tier 3.3 (Le Mans), Tier 4.1 (Monza Road) and Tier 12.3 (Daytona Road). ‚Short‘ appears to be defined as any race with 2 or less laps, irrespective of track length. Here you will find that all the offline bots are one bot lower. So in the example above: Top offline bot is The Raven (-1) and then you find -3, -5, -7, -9, etc.

It works every time, and it doesn‘t matter what the reference bot is, the logic remains.

The last cup in this series (Tier 16.3, Hockenheim Grand Prix) is a bit wierd. Here all the offline bots are five bots lower than the current reference bot. So in the example above: Modesto Lingerfelter (-5), then -7, -9, -11, etc.

There appear to be no bots lower than Moises Gaby, so if the reference bot is quite low, the spacing of the offline bots just compresses at the bottom end as necessary.