I still feel like there should still be some kind of minimum amount of points awarded for wins. It's really not fair to get awarded with "No Change" for a win against an opponent whose level and rank are much higher than mine.
Then it means he had a lower mmr than you. The points are based on the difference in mmr, the bigger the difference the more the total points are added/substracted at the end of the game. Even if he's level 27 rank 9, it just means he went on a losing streak and was behind you on mmr.Darkmor;n8468380 said:I am not sure how they are calculating MMR but that isn't what needs to be changed. What needs to be changed is the ratio they add and subtract Ranking points after the match.
Perfect example just happened: 9th Tier Level 29 vs. me 7th tier Level 18. So I lost one round he won 2. When I get to ranking point distribution screen I lost 72 pts?
I admit I should lose something but I think the ratio should factor in his tier rank and level and vice versa if I would have won I should got a larger amount of ranking points than i would if I was facing someone who was the same tier and rank as I am.
Like I said I don't mind losing to someone better than me, I just don't like losing very, very hard earned Ranking points.
[–]rethazCD PROJEKT RED 135 points 2 months ago*The system does indeed take inspiration from the classic Elo rating system and the assumption you make regarding it are generally correct
As a general rule of thumb it can be said that a higher K-Value is basically always disadvantageous since there many negative things attached, like e.g.: the higher the K-Value the lower the importance of long term performance shortterm performance matters much more which leads to unwished results like: - bad Matchmaking - not finding games because the MMR stretches way more cause of bigger volatility and swings and many more.
This is true. Our K value was working significantly differently from the 19th of December until Mid January. Our simulations predicted drastically different player behavior than what turned out to be reality once the ranking system launched. The top of the leader-board was topping out around 2200 instead of 4200, so in order to allow us to collect more data and also to allow players to have more fun and feel a sense of progression we standardized the K value as a temporary measure. We have learned a lot from both of these periods about player behavior, but also about expectations.The advantage of having a higher K-Value being that you can climb faster with lower amount of games (pretty selfexplicatory if only shorttermperformance matters) => So in general it can be said that you want to minimize this value as much as possible as long as it´s halfway feasable, that means that if that value is extremely low it could lead to the disadvantage of having to play too many games in order to find out your true MMR.We do have a dynamic K-factor in quick play and we are observing the trends and data here still. As mentioned above we also had the same system in place when the ranked system launched, it was present for several weeks. This was not fun for players and they couldn't obtain the rewards we had intended, the system also felt incredibly grindy. I am sure there's a few reading who can attest to that
The K-Factor in Chess once a certain amount of games have been played is "10" at the moment. One game of chess usually takes up many hours and rated matches usually don´t take place on a consecutive basis. The K-Factor in Gwent seems to be around "100" at the moment (I don´t have the specific number but I would guess it´s around that figure.
Given that Gwent games are accessible at any time and take less than 15 minutes/game in average it seems to be quite weird to be the case, since you would expect it to be equal or lower than 10, but def. not 100. Because of that reason we can encounter a lot of problems specifically at the high rank competitive play which would be: - very long queue times, very often not even finding a suitable match within 20 minutes - the MMR not really reflecting your playskill, very dependent on the last 20-50 games you executed which is basically your daily performance.
Our ranked system has two main goals.
- To allow all players to gain a sense of progression and achievement.
- To show who the best Gwent players in the world are.
For the first goal to be achieved the system needs to encourage and allow progression, you purposely lose less points at lower ranks in order to achieve this. This allows players to obtain a decent rank through hard work (decent does not mean rank 15).
For the second point to be true the above seems contradictory. However we actually normalize the points loss / gain as your rank increases, so once your are at the higher ranks the system moves back towards being zero sum and no longer inserts artificial points into the system.
Between how the system was working initially and how it has been working recently we have collected a great deal of data and feedback. And likely the most sensible and logical solution is somewhat of a middle ground.
We still aren't happy with the player progression and we'd actually like the average player to obtain a rating of around 800 high than they currently are. This means more economic rewards for everyone Since this isn't happening yet we will be making some tweaks here.
At the top end of the system it has led to some issues where players are often incentivized to "camp" once achieving a high rating. This is in part why it can become difficult to find a similarly rated opponent. We believe K returning as more of a factor will help with this as punishment and gains would be less swingy towards the top end of the ladder. We are running simulations on this with various weights attached to K, I can't give an ETA on this but it is something that we hope to have in place before the game transitions to Open Beta.Food for thought and hopefully this will see quick change before people get used to climb those MMR points very quickly and feeling bad about "only" receiving let´s say 6 or 7 points in the future.This is very true, but instead of making the person gain less points we want to adapt the system around the concept of the player getting more points and still keeping it balanced. We are aiming to make the system more balanced by modifying the win% expectation formula to work with higher MMR differences.
As a final point, something we will also be adding as a quality of life change for high rated players is a pop-up that asks if you wish to enter extended matchmaking or not. This way it is your choice if you wish to risk your points on a game that might not have the optimal outcomes for you.
We haven't detailed how our season transitions will work yet or even if there will be seasonal transitions. We aren't ready to talk about that just yet
When we are we will be happy to take on board feedback from players. However something we can commit to is that we do not want to destroy the entire progress a player has made while participating in ranked play.
(Please keep in mind this statement does not apply to when the game transitions to Open Beta, the entire game will undergo a full reset, details of the compensation for this are detailed in the following link:http://wpc.4d7d.edgecastcdn.net/004D..._Update_EN.png)
Martin33a;n8532770 said:I suspect that this has been debated many times, but hey.
Martin33a;n8532770 said:I suspect that this has been debated many times, but hey. The Match-Making (and ranking) is a joke. Lvl 32 Rank 12 (me), and up against lvl 70(...) Rank 15. The inevitable loss cost me 63 points of ranking. I beat a guy who was lvl 52, rank 14 (narrowly) and gained 12 points. The losses, whether close or not, cost us huge amounts of points, and wins, especially when close to the next ranking threshold net us sod all. Playing people who are 15-40 levels higher than us may not look much when they are only 1 rank above us but especially at lower levels the difference in resources (ie. Legendary cards) can be huge. Level and Rank need to be matched to make it fair. And the Scoit'ael Passive should be dragged out back and shot. I get the idea of it, but forcing someone to go first when they are a round down, playing Yaevyn (who pisses all over Dinar or Brina) and bleeding them out of cards doesn't make for a fun game at any point for the person on the other end of it. As various streamers have stated, there's no sense of achievement for beating the ST Control deck either, just a feeling of thank f*** that's over. Rant over.
mortimerMarshall;n8832370 said:I think it should be based on the round it took an points won by like if you won you'd get a base amount of points then if you win by more than ten points it gets boosted and so on and so on. That way it promotes winning by large margins.
4RM3D;n8832460 said:Which is a very bad idea because it locks you into playing certain type of decks.
mortimerMarshall;n8832810 said:Not really it would give a more reliable idea of what you'll get/ lose at the end of the game meaning people could be less inclined to rage quit as they'd know what they're losing in terms of points. Also it would be a better incentive for winning securely and if it is close it means people don't loose out too badly.
4RM3D;n8833010 said:None of those points counter my argument that you are locked into playing a certain archetype (to maximize the difference in strength). Thus, still a bad idea. Besides, players will get even more frustrated when the opponent pulls out a Scorch in the final round.
mortimerMarshall;n8833120 said:Well no because you'd still get a decent amount of points just additional if you win by a lot so it wouldn't force you to change style but if you win by more you get a better result which I think is fair. Well that would be a good save by the other guy but still you can predict what you get rather than getting a ridiculous amount of point deducted for a close match.