Letsg0Met586, on 29 March 2015 - 08:05 PM, said:

First off, I really appreciate your thoughts on this! Thinking about game design is as fun as actually playing for me, so I get excited when someone creates a truly intellectual discussion about something like this.

Second, I apologize for using crude guesstimates rather than taking the five extra minutes to do the math correctly. Thank you for taking the time to calculate the actual numbers. It was lazy of me, but it also revealed some underlying disrespect in my personality that I didn't like seeing. I feel it's disrespectful to disagree with someone but be sloppy with one's argument, which is exactly what I did.

Third, I realize that because I used sloppy non-math, it will be harder for any points I might make in the future to be taken seriously, so I won't belabor anything. I'll just say I think it's possible that we're both correct, in a way. I absolutely agree that the more afk players per match, the higher the influence ratio goes for the active players.

However, I also think that using a mean average reflecting that ratio paints an incomplete picture. Even though the ratio goes up as the number of "normal" players goes down, I think that may reflect the many lopsided wins the hero will get--7 kills to 0, or 7-1, or 7-2. When he loses, it will tend to be more close, since he counts for two players. I think this is reflected in your calculations as well: the greater skill differences favor the hero, but there's a significant and growing percentage of unfavorable situations.

In the first two calculations, with 0 and 1 afk players, there are 0% unfavorable situations. This doesn't mean the hero will win 100% of the time, of course. But still, consider that at 2 afk players, there's a 19.2% chance of an unfavorable situation. The hero *might *still win the match if he can play as well as 3 normal players combined, but it will be much harder. At 4 afk players, there was a 7% chance of a badly unfavorable situation, but also a 36.7% chance of two equal teams. We have to assume that the hero will only win half of those, of course, and they are only equal because the hero counts for 2 players. His teammates are still worse than the other team, only he makes up for it half of the time. In this scenario, an additional 18.35% of the losses he takes are the fault of his team. After all, he played as well as 2 normal players, but in a close match his teammates let him down and he lost. That makes a slightly higher chance--25.35%--of a frustrating loss caused by his afk teammates. The pattern is rough, since we're using only 3 levels--afk, normal, and awesome--of skill.

Since there are many different "levels" of poor players, I think the pattern does even out as you add a more highly-varied mix of good, decent, bad, awful, and afk players. For example, a truly legendary player who is worth 3 "normal" players would likely win every single match if his teammates were as skilled as the enemy players. That's a 9-7 differential. But move down to 3 afk players on his team, 0 on the other, and suddenly it's a 6-7 disadvantage. His win rate moves down by 7%, since it was practically at 100% before.

Or, a "pretty good player" might be worth 1.5 normal players, and then of course would be far less likely to carry his team at a disadvantage. Or, the "legendary" player might be the bottom tier; with two evenly-matched teams, he would still be able to sway the balance toward winning, but give just 1 useless player to his team and suddenly his chances of carrying go way down.

So to sum it up, I don't think a higher mean average skill differential ratio necessarily translates to a higher win rate, although it might. Also, in retrospect, I seem to be most concerned with what percentage of matches are a true disadvantage for a good player. That doesn't necessarily translate to a lower win rate either, but it very well could (example being the non-existent "legendary" player who only loses if his team has 3 afk players on it).

**Edited by NevirSayDie, 30 March 2015 - 03:50 AM.**