Do you have any suggestions regarding the rating system?Most of the participants are OK with the current system. Many players
really liked it.
But you guys had
A LOT of interesting answers. Some of the most interesting ones:
- Yotoon! was experimenting with TrueSkill (based on elo ranking), other rankings should be investigated.
- Perhaps decrease the difference between the top places.
- We will later propose our new variants (an uN*DeaD player).
- We need a smaller point step between top places.
- uN*DeaD|Enter's adaptation of Nascar system
- Delete Zerg from this world and you won't need this question in the Survey
- Number of points should be proportional to the round difficulty - interesting thought, but it is arguable. Critics would say that organizers give more points for, say, a plasma run, because they support Dex who is good at plasma, or they give more points for a strafe map to support a strafer etc. Difficulty is subjective, so it is impossible to come up with factors that suite everyone. But we should think more about it.
- Take time difference to account and add to points (somehow). So that 1st place 30:000 and 2nd place with 30:008 are less different in points than 1st 30:000 and 2nd 31:000.
- Add point penalties for skipped maps. aDr played half of the maps and ranked high. But other players who player all the maps could not catch him up because of big steps between places and no round skip penalty.
- Anything but this system.
- Zerg and HOX proposed good systems in the final round stream.
- xas/indexfan time bonus scoring idea is interesting.
Some people proposed too sketchy and inappropriate systems, like:
- Number of players minus rank - that's a linear function and it doesn't reward the winner. Top-1 player gets 167 points, and top-2 player gets 166? Seems unreasonable for the winner to play hard just for 1 point difference. The current system makes a 200 points step, which may be too much. We need a more linear distribution, but not quite linear.
- The lowest average rank should win - this is even worse, because it doesn't let one skip a round. We all have real lifes and may have emergencies etc. Skipping a round should probably be punished. But lowest average rank means no rank in a skipped round, no way to calculate the average - disqualification? Point-based systems are more appealing in this regard.
- Elimination-based systems - This makes it a competition on one map. Regardles of ranks in the 1st rounds, those who pass qualification of penultimate round will find the "winner" in only the last round. It makes no sense. No motivation for them to play hard in early rounds, and no way for average players to play in later rounds. We want as many players to play as possible. We want to have fresh blood in DeFRaG. Everybody wants to get maximum of a world competition that happens so seldom.
- Sum of time - this system was used in 2008, and a quote from another player: "in 2008 dex won 7 maps out of 9 but lost the competition because he lost too much time on one map, which makes no sense."
- Exponential function y=a^x - won't do, because most of players didn't like too big point gaps and asked for a more linear function.
In the end I make the following notes for myself or for other organizers of the next DFWC:
- Consider a more linear system with smaller point steps between the leaders, for example Enter's adaptation of Nascar?
- Consider multipliers for different maps based on difficulty or proximity to the final.
- Consider adding/deducting
some points based on the time gap (xas/indexfan to take a look at).
- Consider a penalty for not sending a demo for a round.
If we implement all of that, will it not be too difficult to understand? When looking at the round results table, will players not complain that they don't understand why they got less points? Also, all these features should be tested on results of previous DFWCs.