Xtreme Gaming Network

The Ballast Thread

Author Topic: The Ballast Thread  (Read 4465 times)

Offline Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 11033
  • Karma: 152
  • AC Admin
    • View Profile
The Ballast Thread
« on: August 02, 2015, 12:07:43 AM »
Here's how the ballast worked, to good effect, in our old netKar PRO league, fro your information. This is what I'll be doing in AC as well, when we decide to use ballast. This is a proven technique that worked well for years in our netKar PRO league. We won't necessarily be using ballast every season.

The important stuff

In a nutshell, the faster you are in a race, the next race you will carry more ballast to slow you down. If you're not so fast, you might lose ballast. Your ballast changes from race to race (not round to round). The more often you adjust the ballast, the more accurate it becomes.

At the end of each race, my program works out the average fastest race lap time for each racer, who was within 6% of the fastest lap time (this excludes lap times of people who are too slow, so that everyone else doesn't get masses of ballast piled on). Basically, the faster your best lap time is compared to the average, the more ballast you get. The slower your best lap time is compared to the average, the less ballast you get.

How much ballast?

The amount of ballast you need to achieve a certain percentage slow down in lap time varies from car to car (as cars are different weights). This base amount gets set by experimentation before round 1 of a season and then gets refined each round as we see how different people are slowed down, on average, by the ballast in practice. Some people will be slowed down more by the same amount of ballast; some less, due to driving style. It all gets average out.

Bonus points

The ballast tightens up the field, but people who are carrying more ballast get bonus points, e.g. 0.5 points for every 25kg of ballast. This way, faster people are still rewarded on the leader board.

The fine print

If a racer didn't get to complete 4 laps or more in a race, their ballast won't change for the next race, because they haven't done enough laps to calculate a decent average. This avoids unexpected skewing of your ballast.

In the first two rounds, your ballast can jump up and down a bit before it settles on an average. To "smooth out" ballast changes after the first two rounds, your ballast changes are reduced to 70% of the calculated ballast change. This stops your ballast from jumping too dramatically up or down. E.g. if it's calculated that you need 100kg of extra ballast after the first two rounds, the program will only give you 70% of this i.e. 70kg. Likewise for decreases.

An example

Here's what happens in my program after a race:



The average fastest race lap was 1:47.639. Looking at Guybrush's line, his best time was 1:45.691, which was 1.948 seconds faster than the average. As a result, he gets 46kg of extra ballast for the next race (he started on 0).

The people like Rob who were slower than the average (0.272 seconds slower, in Rob's case), lose some ballast from their starting weight.

“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Offline Guybrush Threepwood

  • Hero Member
  • *****
  • Posts: 2133
  • Karma: 69
    • View Profile
Re: The Ballast Thread
« Reply #1 on: August 02, 2015, 06:13:36 AM »
Sounds like you have it all sorted Wally.

Would it be worth looking at the average lap time for each drivers fastest 3 or 4 laps, or a % of laps?  Possibly not as I guess the more laps you use, the more chance you will include inaccurate data such as being held up/track evolution or something.  It would be interesting to see how the above example would change.

Offline marty

  • Hero Member
  • *****
  • Posts: 3747
  • Karma: -63
    • View Profile
Re: The Ballast Thread
« Reply #2 on: August 02, 2015, 08:05:53 AM »
Using the example there it seema it may be better calculating ballast off the slowest guy. For example I was about 1 sec quicker then Matthew and I get 25kg more but then you have 2 seconds between Wally and chap yet they both have the same weight.

So looking at it after this race between Wally and me with a time gap of 2.4 seconds there is 50kg but then the 2 seconds behind him those guys get no help. This way it will just bring the top half back a little towards the mid pack pace. The back of field would still be the same.

Wpuld it not work better if say the fastest guy hot 100kg and the slowest one 0. From further races it may balance out better but as there is no way the guys in the back can lose weight to help them maybe just race 1 should use ballast off the last guy then each other race could use the system as you have already.

Offline Wally

  • Administrator
  • Hero Member
  • *****
  • Posts: 11033
  • Karma: 152
  • AC Admin
    • View Profile
Re: The Ballast Thread
« Reply #3 on: August 02, 2015, 02:52:24 PM »
What that example doesn't show is how it pans out after a few races... this would be the first race where everyone starts on zero. It's a bit of a bad example because you don't see anybody's ballast going down. What happens is most people end up with some ballast, and only the slowest guys are on 0kg. By using the average lap time, this makes some people's ballast go up and some go down.
“You can please some of the people all of the time, you can please all of the people some of the time, but you can’t please all of the people all of the time”

Offline Freezer

  • RestrictedRacing
  • Hero Member
  • *
  • Posts: 1296
  • Karma: 33
    • View Profile
Re: The Ballast Thread
« Reply #4 on: August 02, 2015, 05:46:43 PM »
What that example doesn't show is how it pans out after a few races... this would be the first race where everyone starts on zero. It's a bit of a bad example because you don't see anybody's ballast going down. What happens is most people end up with some ballast, and only the slowest guys are on 0kg. By using the average lap time, this makes some people's ballast go up and some go down.
This is good news . . it did work well in the Nkpro days and should mix up the results a bit.  Looking forward to it.

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal