The new update and the changes done so far are going in the right direction, but some efforts are still required to make a good game. Here are some my observations after playing about 100 matches in ranked and some proposals how i would be possible to improve this system:
1. Unbalanced teams.
It is faster than before, but actually you can get almost anything. The teams are unbalanced on the both, on the profession and on the skill level. For example, one time I got a team of 4 rangers or another time 3 necros. The results of such teams were disastrous. Some people cannot properly play, have probably just started to play at all or have played before another profession. The last probably results from the ugly ranking system that is blind about the profession.
Proposal > The issue with unbalanced professions could be easily solved by just giving people an option to select what kind of a team they would like to receive. At the best it would also require to introduce the rating person/profession.
2. Unfair measurement of individual skills on team-based results.
There is still a huge discrepancy between the team-based result and individual ranking. With the recent update it seems that the required individual statistics of players vs. its own team are available, but the reward still heavily depends on the end result of the whole match. It is also the major source of all hate, impoliteness and flame that we have in PvP now.
Proposal > Why not just to value the results of randomly-build teams and party teams differently? For random teams the results could be measured individually heavily based on the individual results vs. own team, and for a party all together as one individual. Something toward a principle „What You Giving Is What You Get“. So, for every game players received some positive number of points. The total rating could be calculated as an aggregated total number of points divided on the total number of games played.
This would allow measuring the results for parties of any sizes. It is clear that a random party anyway is not competitive with a full organized party, but why not to rank all these matches separately? It would be reasonable to see at least 3 types of ranked matches: duels (1 vs. 1), teams random individuals, full fixed parties or guild parties. Actually, the guilds have already such teams, why not to use them for ranking?? Here I mean only full teams.
3. No fair treatment for disconnecting, afk or just not playing team members.
May be every 5th match I have had such people in the team. A match of 4 vs 5, assuming that all players have similar skills will be lost for the first team almost with sure, but it will be ranked and there are no any really working mechanisms implemented prohibiting to do such things. Sometime people are leaving a match after some minutes or just staying inactive/active in the home zone that almost inevitable follows to a lost match.
Proposal > Matches with disconnected, afk or not active people should be terminated immediately as this happens. The person that provoked the termination should receive a progressing decay in the ranking equal to an average loss, the other persons should be ranked individually according to the individual statistic vs. team that was achieved so far.
4. Professions are not taken into account.
As it was mentioned before somebody who has mastered one profession may be not the best player with another. Currently the rating system does not count this anyhow.
I would also like to see the whole statistics of the best players together with the played professions, not just only on players. I am sure that for the most players it would be also interesting to see which profession is giving the best results. Or, you just afraid to show these results, maybe because of the obvious unbalance of particular professions?
Proposal > Calculate, use and present the rank per person/profession.
5. 50/50 averaging of the individual results playing in random teams.
If the matchmaking system giving us every time almost equally skilled teams of random players, then every time the probability for a win or loss for individual people will be about 50%. “Statistically speaking” this means that independent how good or bad you are playing, from 100 games played about 50 will be lost. Adding here the factor of afk and wrongly matched people receive that more than 50 games will be lost. As long as the end results are still having significant impact on the rank, this system will naturally press all ratings to the low middle of the scale and, therefore, will not show a real performance.
Currently this law will be partially mitigated by people playing always in a good duo team. In this case the probability that your team will win increases and the probability to receive afk people in your own team decreases.
Proposal > The solution is already presented in above points 1-4.
(edited by Kris.4603)