We'd like to discuss the ELO system with you.
Over time, this particular game mode has undergone several changes that have largely proved to be successful. So this is where we are currently at:
- We’ve reached the point where there are almost 1,000 cards available in the format and in which "optimized" decks exist by mixing up clans. It would seem that "historic" cards (ones that are several years’ old) have an unchallenged place in decks that work well in ELO and, as a result, are greatly limiting the emergence of new deck types.
- The format has dramatically improved with the arrival of new clans and particularly new cards which mean that decks can be significantly improved (it is now possible to produce a proficient deck from each clan). By this, we’re not referring to stomp decks used to race to the top of the rankings, but rather to decks that work well in the ELO mode.
- More effective cards per clan also means greater possibilities of monoclan domination.
Our current aim is to give the ELO mode room to expand and evolve by adding a criterion to staff bans.
So this is what we propose to breathe new life into the ELO mode:
The age of cards will become an important factor in what we decide to ban from the ELO mode. The older a card is, the more it runs the risk of being banned. This will enable us to dislodge certain cards that have become fixtures in ELO decks and, in doing so, make way for new cards to emerge.
We will select between 0 and 4 cards per clan (at each wave of bans) which we can ban based on this criterion.
For example, let’s take a look at an All Stars deck that is currently very effective in ELO. Marina (2007), Jessie (2008 ) and Randy (2008 ) very often have a place in their star category. If we ban Marina and Randy, Nathan might find his place or would at least have the chance of doing so. In the same way, the presence of Uranus partly prevents the creation of a new reducer for the Sakrohm clan because of his dominance. Several cards may not seem worth considering because of their direct competition within the clan. However, we believe they could provide interesting alternatives (I'm thinking here of Leviatonn VS Klawz or Stalfhaust).
This new criterion of play would allow us to better manage the game's evolution. We would fix dates for new bans every X months, of course taking the time to first discuss this with the community, as we did with the last wave of bans.
Banned "historic" cards would then be definitively expelled from the ELO mode.
In short this new system would allow us to:
- Refresh the ELO mode on a regular basis by providing genuine renewal to the metagame.
- Give new character cards a chance to make their mark on the game mode.
- Allow new deck types to emerge by combining the various clans without having to play/challenge certain unavoidable cards (by getting rid of “historic” cards, players would have more scope for creativity).
- Reduce the power of monoclans in the game mode.
Initially, each clan would not be affected in the same way but over time a real change in the mode’s playstyle could emerge.
Ultimately we feel these suggested changes will provide ELO devotees with yet more scope for discovery and enjoyment.
These are our proposals. We very much look forward to hearing any comments or suggestions you may have on them.
edited by Espectroscuro monday 23/09/2013, 16:32
"I'm wondering where the strategy went"
It went the same gutter as anything with static systems. You can ban anything from ELO and new STATIC system emerges to replace the strong and weak points with new strong and weak points. The fundamental problem is that there is always "optimal solution for problem" and banning characters does not change that, there will be optimal solution for problem after the bans too.
To make it more interesting and appealing you should make the problem more complex, when you ban characters nature of the problem does not change (find new optimal configuration from possible choices. Thats just average task).
By giving "ELO cost" to characters and changing that cost dynamically problem gets more complex. You have to mentally measure the cost against true card usability instead of simply picking the best of the available choices. There would be "optimal solution" for this too, but its formulation is much more complex than in "unbanned free picks" scenario. The problemsolving is not impossible or even hard, with intuition you can build strong decks without them being optimal, thus it becomes interesting area for competitive play - instead of having few "proofed" solutions you'd have many strong alternatives and balance that changes weekly.
Stillborn, I think many are confucing "ELO points" with "ELO score" from game mode itself. "ELO price" is more describing word for the concept.
Generally speaking, "Star cost" is these days really bad measure of cards usability. We really need some other way to measure cards real potential and "ELO price" would fill this role about perfectly.
This comes out clearly when we measure something like Mawpin against Gary in ELO setting. Starcount doesnt mean anything, Mawpin would have ELO price of about 150 points (good low star card) while Gary's price is flat 0 (waste of deckspace).
@DeepEnd - : Creating a complex system of points to make a deck for ELO or to ban cards is really complicate, and more for the players who are not expert in ELO and this rather will be annoying them (if they have to calculate each value of cards and, what could be the real value of a card?). If you lose the player while he’s only creating a deck it’s not worth it, I think.
Not too complicated and the concept is easy to understand. Simple tool to show only cards that are "within the budget" and visually seeing the prices is enough clues for players to know what they are supposed to do, and having some "Elo cost" label for decks ELO price. People are not stupid and if they are they should not play ELO in the first place
And as far as developement goes there is no real technical or code barriers that forbid this thing from happening (unless its matter of developement costs vs. willingness to change the game). UR has created far more complex patterns in past. If UR is willing to commit then this thing is possible - you wanted opinions and ideas - you got them.
The problem you are presenting about ELO requires a bit more complex solution than is possible with the immediate tools available. If you dont plan to do it properly then dont do it all.
@Agree with deep end:
This is in fact THE solution to the problem you presented originally. It's also a way to give cards with no meaning within the current system some meaning. You can't make cards that will replace staples unless you make them stronger or fill holes that are there for a reason - once everybody has everything clan bonuses rule. The vast majority of cards you print or can print are pointless by design if they compete with a staple. Look at Junkz 5*, all playable but only one has 8 damage and SoA. You can print Junkz 5* for three years every release, unless one of them is a 7/8 SoA it's highly unlikely anyone will put one ih his or her deck if they can have Fuzz. You've banned a huge ammount of stuff and people just filled the holes with the next best things and ELO is very static. I discover an underused card every time I have a streak, but it ultimately shows why it's underused - something that's there all the time is just more solid.
It's also why you get comments like "Clan X is good because it has card X". Berserk isn't good because it has Spyke and bad if he's banned, but the system is forcing the people to look at the game through this perspective all the time. The most difficult thing in the game is learning which cards are the absolute best in their clans, which cards make clan X playable. Apart from the played cards almost every card that's not them is a newbie trap and a distraction, decent cards are commented upon as "good untill you get X".
Furthermore, I could right now earn a few million clintz if I simply sold all the cards I will never use - because as long as i have the staples for each clan I can always get my 5 or 10 ELO credits. My collection is worth 4-5 million or so, and maybe only a million of those are ever going to be used and are worth having. The only time I'd miss them in practice is when I need something for the ld Missions. Just think about that for a second.
And sorry for 3 posts but what it mens that the only reason to have the vast majority of cards beside the very best is for the missions, and the chance they go Cr. If I sold everything I'd have enough money to buy new cards every time you make a new one that completely obsoletes an old one, or you ban one and another becomes the absolute best choice. This would last me several years and I wouldn't have to waste a dime on buying credits.
Why not spend the money on something? On what? Old expensive Cr's can't be played anywhere, and besides the staples there is nothing of any real worth whatsoever - I could just play the market for more clintz that I don't need...
Perhaps by polaying with how many cards a deck need to involve you can get some change to things. It will be a bit harder but for example requiring 12 cards instead of 8 will make people go more into dual clans and give place for some new staples. Going to 16 cards and 50 stars should would be a bit too big of a change but something can be tested I guess.
But bigger decks will give room for more new cards all in all as when you need 8 cards to make a deck there are a lot of obsolete cards. If you need 16 cards it allows for 2 times bigger cards collection. however the problem with this game and adding stars + bigger decks is that there will be more situations where a 2 2 2 2 star hand meets a 4 4 5 5 hand and will not be too much strategy involved anymore. but perhaps going small to 10 cards and 30 stars might do something ?
Actually as IT systems designer background I can lay out the requirements for such scoring system if neccesary. It might require some internal changes to codebase but those are basic stuff (compare value of deck to max price? complicated? Nope. Add additional database row to card stats for its "ELO price". Complicated? No.). Only real challenge may be gathering the data from actual games from ELO for analyzations (if such detailed logging feature doesnt exist already) but that too is matter of storing values to database and then in the end of the week look out for cards that are used a lot and are victorious in game - thats some data mining but basically any average programmer can program that task. Then in the server-client framework put the check if users deck can be used in ELO and youre pretty much done since the actual gameplay doesnt change. Except for the deck building UI changes.... but since you are able to do completely new UR beta from almost scratch I'd assume UI work to present few values and add few more search criterias to cards & provide help page is not a big thing.
There is work and changes related, but even the most advanced stuff of this thing is not harder than average task. Most work should go into planning this thing out but you have already rough edges on this thread, concept-proofed in meta level by vigilant minds of community of subject related professionals.