by JR Cook - 8 years ago     show comments

Last week, we started looking for options to develop a better ranked system. We figured out all the key points, so now its time to develop a finished product.

The Power of a Name

Our new mode will be called Legends Mode, it just fits with the lore of Hearthstone and it shouldn’t make newer players feel bad about it. Everyone knows how hard/easy to get legend is, right? Still, this new mode is aimed at those top 1% players that actually make legend every month. So it needs to be named as something special, you can’t just name it ranked, and change the current ladder to something else. Naming is important, and we need new players to feel comfortable. I’d even go as far as to keep this Legends Mode hidden until you actually reach legend.

Laddering Up

The Legend Mode should divide its players in several divisions, at the end of the month, the top 20% of players in the division should move to the next tier, while the bottom 20% should drop down one.  So we would have something like the typical: Bronze < Silver < Gold < Platinum <Masters < Grandmasters. But how do you determine who is the top and bottom percentage? Well, we need a sort of scoring system, and the most commonly used are based in sorting players by their Matchmaking Rating (MMR).

Matchmaking

Hearthstone has a lot of luck involved. Sure,  skilled players will usually be more consistent and get higher win-rates over a high number of games, but a lesser player can easily steal a win from even the world champion. This is why a traditional MMR system wouldn’t work very well in Hearthstone. Losing a bunch of rating just because a low-tier player got lucky on you is frustrating; a lesser player getting a big rating boost for defeating a proven master, can also throw the system off. Your MMR should be a measure of your skill, the only true measure of skill in Hearthstone is consistency. As we said last week, win-rates are the best way to reflect that.

For match making, the game should try to find a player inside your division that has had a similar win-rate than you over the last 100 or so games. As the time passes, the pool of eligible players should be growing, allowing you to match against players whose win-rates are more and more separated from yours. Should no player in your division be available, then the eligible opponent pool would grow to include the top % players of the division below you, and the bottom % players of that above you.

So your MMR would essentially be your win-rate. But I know what you are thinking, it wouldn’t be fair to make that your score. You’d benefit those who where consistently matched up against worse opponents. So we need a correction factor, and I believe that some form of Strength of Schedule (SOS), would actually help out here.

Scoring

SOS is basically a measure of how good your opposition was, in this case it would be the average win-rate of all your opponents. The key here, is that for scoring purposes you will calculate your SOS at the end of the month. It doesn’t matter what the win-rate of your opponent was at the time of your match, but rather how well he did the entire season. The high variance of Hearthstone means that the only way to accurately assess performance is over a large number of matches, this type of scoring system would follow that principle. Provided that you complete a required amount of Legend Mode Games during the season, your score would be looking something like this:

Your Score = [Your Winrate * (0.75 + SOS/2 ) ] * 10000

This would mean that if your win-rate was 57%, and the average win-rate of your opponents was 52%, your score would be:

[0.57 * (0.75 + 0.26) ] * 1000 = 5757

If you feel like the SOS isn’t doing much of a difference, you can try changing the weights around.

Your Score = [Your Winrate * (0.50+ Your SOS ) ] * 10000

[0.57 * (0.50 + 0.52) ] * 10000 = 5814

The only (and perhaps big) problem with this is that your score would change as all your opponents win-rate keeps on changing. To address this, I’d create a cut off point a few days before the season ends. The number of required games should be completed BEFORE this cutoff point, the games  after this point would still count for your winrate, but wouldn’t count for your opponents’ SOS.  So your SOS score would be:

SOS = Average of ALL of your season’s opponents win-rate on THEIR matches between start of the season and season cutoff point.

Is this intuitive? Not at all. Is it a bit overly complicated? For sure. But this is a system that is designed for the elitest of the elite, this are the people who are trying to become career players, they have the time to analyze and understand these formulas.

But the important thing is that this means that after the cutoff point, only your matches will affect your final score. This means that players will only depend on what they do on those last days, and what the people who are right below and above you do. There’s a still issue to tackle, how do you know if you are safely on the top percentage of players? How do you know if you have a big enough lead to sit those last few days, or if you have to put in the time to grind a bit more?

Freedom of Information

When you have a system like this, allowing the players to have access to this important information is key. The scores of each players, the leaderboards, and even the match histories of all players should be readily available. Just look at Starcraft profiles, that’s exactly what competitive Hearthstone needs.

How pretty would something like that look on a Hearthstone skin? In any case, this would be key for players as they’d be able to assess whether they need to keep playing, or if they have already secured their spot. There’s absolutely no need for games on the latter part of the season to be worth more than those on the beginning, right?

Hearthstone World Championship Tour

Ok, we’ve covered matchmaking, scoring, league progression, and accessibility to information. But how about qualifying for the World Championships? Easy enough, what about this:

Bottom 20% of Bronze League = 0 Points.

Would Start next Season on Silver League = 1 Point.

Would Start next Season on Gold League = 2 Points.

Would Start next Season on Platinum (Truesilver?) League = 3 Point.

Would Start next Season on Masters League  = 4 Points.

Would Start next Season on Grandmasters League = 8 Points.

Top 20% of Grandmasters League = 15 Points.

There are a few fatal flaws with this system though, but it would take way too long to explain and correct them. That’s why next week we will take an in-depth look at how this ranked system shuold be implemented in regards to awarding Hearthstone World Championship Tour points.

See you next week!