Aresye, on 09 November 2014 - 02:07 PM, said:
So using the probability formula posted on that thread and plugging in a 100 Elo difference, the probability of the lower Elo beating the higher Elo comes out to 29%. Or a 71% chance of the higher Elo beating the lower Elo.
On the higher end of the spectrum with a 120 Elo difference, the probability of the lower Elo beating the higher Elo comes out to 25% (75% chance of the higher Elo winning).
So technically, the Clan side should have won between 71-75% of games with an Elo separation between 100-120, which means...
...the IS side was actually performing better and winning more matches than predicted.
ggthread
The IS needs to be winning ~50%. In fact a 45/55 spread is going to be problematic over time given the 'collect tokens' approach to taking planets.
You're not going to have a situation where you've got more IS players and better IS players than Clan players. To do that you'd have to control player behavior, which isn't going to happen.
There will be more Clan players and the most competitive players will gravitate to Clans, all because Clan tech is superior than IS tech (even if it's just a few mechs, those 3 or 4 mechs can fill the roles of 20 IS mechs via omnipods). So you're going to have the Clans rolling the IS.
The result of that will be an exodus of IS players from IS vs Clan play. You'll have Joe and a handful of like-minded folk getting rolled again and again and again while most the IS play is IS vs IS. Any sort of matchmaking will fail and fail completely due to insufficiently deep populations.
You're not going to get any significant group of people to happily go grind out matches that they lose 2/3rds of.
If Clan vs IS tech is not absolutely 50/50 or within a couple % of that then CW is doomed before it starts. It's a concept based on results averaged over time. If after 100 tokens have been earned and one side, even if player skill/population distribution was perfectly homogeneous between the sides, is going to win 55 of them based on tech advantage and the other 45 due to being at an overall disadvantage you're dooming the whole process.
Quirks are a good start. CW without IS/Clan balance will fail. To pretend that people will behave contrary to human behavior and their own interests
overall as a population average is not reasonable or rational. Saying 'yes, well, Clans should win most the time' is all well and good - until it happens and suddenly most people either only play Clans or only play IS vs IS. Then what do the Clans do?
They play Clan vs Clan or they wait 30 minutes for their turn to stomp on Joseph Mallan, two other people who just love the IS vs Clan dynamic even when they lose 2/3rds of their matches, and 9 nubs who decide the game sucks and will never play it again.
A 100 pt difference is a 25/75 prediction. A 36/64 would be about a 71 pt variation in Elo. They may have a prediction variable based on relative tonnage, gleaned from historical data. Wouldn't surprise me.
No crystal ball or complex math required to say what happens if CW isn't pretty perfectly balanced for IS vs Clans though.