Queadah wrote:potterman28wxcv wrote:gaining rating based on the amount of time you actually spent in the race.
Coming to think of it, the other way round might be more balanced:
- losing points gradually based on the amount of time you didn't spend
This way you don't touch the points earned by drivers that played all the event through (no introduction of yet another point balancing issue), and you affect only the "faulty" players.
Being first is something like 250pts. 1 lap is approx 1min, so make the quitter's score his total plus what the bot earned minus 250pt/min-not-played
So the more he doesn't play, the less rating he loses ?
If he doesn't play 1 minute, he loses 250 pt
If he doesn't play 3 minutes, he loses ~80 pt
Doesn't make much sense. Something the likes of 250 * (min_not_played / total_race_time) would fit more your description.
By the way, on my model, I don't touch the points earned by the drivers at all. Mathematically, this would be the general formula for gaining rating:
rating_gained = match_rating_win * (amount_of_time_you_spent / total_time)
Where match_rating would be based on the rating of the opponents you faced, along with which one of them you defeated.
If someone plays for 100% of the match duration, he gets full rating. If someone plays only 25%, he gets quarter rating. So it would be disadvantageous for someone to play only 25%, since all the others would get 4 times more rating than him.
The rating lost would not depend on the amount of time you spent, so it would be:
rating_lost = match_rating_lose
where match_rating_lose would be based on the rating of your opponents, along with which one defeated you
In the end, your rating would be increased by rating_gain, and decreased by rating_lost : new_rating := old_rating + rating_gain - rating_lost
So what happens if someone joins a race and leaves right after joining ?
rating_gain = 0 (even if match_rating_win is > 0 because his bot might have gained some positions, he didn't spent anytime in the race)
rating_lost = something varying between 0 and high number. Basically, if he gets lucky, his bot gets first and he doesn't lose anything. If he doesn't, well he loses against everyone.
In any case it would not be beneficial to ragequit in terms of rating. Which is the only reason why people ragequits in the first place: to get rank points faster. If you deny them that advantage, no more reason for them to ragequit.