CoC BB - Matchmaker...how DOES it work?
One of the biggest conundrums in the Builder Base game is the matchmaker. When you're new to the game, you might find yourself very surprised when suddenly you are being matched against stronger players - consistently. I mean...how can my BH6 game be constantly matched up with BH7 games?
There are a lot of little things that factor into the match-ups, but the actual match itself is determine simply by where you sit in the trophy rankings. If you are a BH5 at 2500 trophies and there's a BH7 at 2500 trophies, and you're both online playing at the same time...well...you two might just end up matched to each other. The BH level itself simply isn't a factor.
So, the REAL question we need to ask ourselves - how do I end up with a trophy ranking that makes ME competitive with a BH7 when I'm only a BH5?!
THAT is a more complicated answer.
SuperCell has told us that the ranking system used on the Builder Base is their own version of the "Elo Rating System".
The Elo rating system is a method for calculating performance of players in a 1v1 competitive game. It was originally designed for chess and is named after its creator, Arpad Elo. The idea is to rate player skill levels based on how they perform in previous matches. This rating increases and decreases depending upon how they continue to perform. If you lose a match, you lose points from your rating; win and you gain points.
There are a lot of games out there that have implemented a form of this rating system, including professional sports, video games, and even board games. A little known fact - the dating service, "Tinder", uses a version of the Elo rating system to rate participants in the service.
There are all kinds of formulas that go into the actual scoring on each match. They're not very complicated and are widely known for chess. (See the Wikipedia article linked above.) As I said, though, those who have adopted the Elo rating system for other games have implemented it more in concept than doing so point by point. SuperCell appears to be no different.
We may not know the formulas used by SuperCell on the Builder Base, but we can certainly see them in the way matches are scored. For example - Elo awards more points to a competitor who wins a match against a stronger player, and subtracts more from a player who loses to a weaker competitor. Ever notice that you might lose 29 points in a match but win 31 in the next, then only win 29 in another? One notable difference - in the chess implementation of Elo, a tie results in additional points for both players; on the Builder Base, a tie results in zero points.
In a perfect world, the actual skills and performance of the player base is roughly normal. That's a statistics term where the distribution of data conforms to a very specific set of break points. It has to do with standard deviations, mean/median/mode, and certain percentages of players being within a certain range (1 deviation) about the mean. In a normal distribution of players' performance, right about 68% of all data points will fall within +/- 1 deviation of the overall average. (In other words, the vast majority of the players will fall pretty close to the middle while stronger and weaker players work their way to the sides.)
If the data distribution is normal, then the Elo rating system is a pretty good predictor of match outcomes. As players advance their rating, they will eventually begin to match stronger players who are more on their level. Eventually, the probability of a win becomes 50:50 and that ends up being the win:loss ratio.
Within the Builder Base, we see this constantly. The number of trophies you have is effectively your Elo rating. As you win matches, you get more trophies; lose and you drop trophies. You have the ability to chose your pool of opponents by dropping trophies into a lower trophy range, but this still works within the system because you are reducing your rating artificially to go against weaker opponents. Of course, if you simply play the game as designed, you will gain trophies until you reach a sweet spot consistent with the maturity of your game and your skills as a player.
I've spoken of the Four Pillars of Clash before - Offensive Capabilities, Attacking Skills, Defensive Capabilities, and Base Design. The more players match up across these four categories, the closer their trophy rankings will be to one another. While one can compensate to a degree by being stronger in one of the four pillars, that will only get you so far; eventually, someone who is more balanced will beat you and your progress through the trophy rankings will stall or drop.
To provide a more extreme example.... It is possible to go from a brand new game all the way to BH7 in just over a month. Such a player has all the toys, but his defenses will be low - typically L3 or under. His troops will also be limited - he may have one or two troops that he maxes along the way, but that does not provide a versatile attacking position. If this is his only game, he will also be lacking in attacking skills by sheer virtue of not having had much opportunity to attack.
That means he has all the toys of a BH7 but he's lacking in at least two, if not three, of the four pillars. Would you really consider this player BH7? The reality is that he's too weak to be an effective competitor in the upper trophy rankings and will likely be in the same trophy range as a BH4 or BH5 player who is taking the time to better develop his game. I've had plenty of lower level matches where my BH4 or BH5 game comes up against one of these BH7 games and not only wins the match but even gets a 3-star win against the game.
In other words - my lowly BH4 and BH5 games were performing at the same level as a hyper-rushed BH7 player, so we drew what looked like an underdog match on the surface but was actually an even match because we each resided in the same trophy range (had a similar Elo-type rating).
Generally speaking, the differences between two players are not this stark. It tends to be just one level of BH and defenses are a hodge-podge of levels right around the level of the lower BH competitor. The take away here: BH level isn't always a good predictor of the outcome of a match; you're better at the game than your BH level suggests.
The rating system enables you to fight against peers - opponents who perform similarly against other players of the game. Their skills and capabilities, overall, will be a close match to your own. If you find yourself in an underdog match, know that the other players is almost certainly weaker in one of the four pillars where you're stronger and that you DO stand a chance of winning because his/her performance is right about even with your own.
Here's this article's companion video. I'll be back soon with more...!
There are a lot of little things that factor into the match-ups, but the actual match itself is determine simply by where you sit in the trophy rankings. If you are a BH5 at 2500 trophies and there's a BH7 at 2500 trophies, and you're both online playing at the same time...well...you two might just end up matched to each other. The BH level itself simply isn't a factor.
So, the REAL question we need to ask ourselves - how do I end up with a trophy ranking that makes ME competitive with a BH7 when I'm only a BH5?!
THAT is a more complicated answer.
SuperCell has told us that the ranking system used on the Builder Base is their own version of the "Elo Rating System".
The Elo rating system is a method for calculating performance of players in a 1v1 competitive game. It was originally designed for chess and is named after its creator, Arpad Elo. The idea is to rate player skill levels based on how they perform in previous matches. This rating increases and decreases depending upon how they continue to perform. If you lose a match, you lose points from your rating; win and you gain points.
There are a lot of games out there that have implemented a form of this rating system, including professional sports, video games, and even board games. A little known fact - the dating service, "Tinder", uses a version of the Elo rating system to rate participants in the service.
There are all kinds of formulas that go into the actual scoring on each match. They're not very complicated and are widely known for chess. (See the Wikipedia article linked above.) As I said, though, those who have adopted the Elo rating system for other games have implemented it more in concept than doing so point by point. SuperCell appears to be no different.
We may not know the formulas used by SuperCell on the Builder Base, but we can certainly see them in the way matches are scored. For example - Elo awards more points to a competitor who wins a match against a stronger player, and subtracts more from a player who loses to a weaker competitor. Ever notice that you might lose 29 points in a match but win 31 in the next, then only win 29 in another? One notable difference - in the chess implementation of Elo, a tie results in additional points for both players; on the Builder Base, a tie results in zero points.
In a perfect world, the actual skills and performance of the player base is roughly normal. That's a statistics term where the distribution of data conforms to a very specific set of break points. It has to do with standard deviations, mean/median/mode, and certain percentages of players being within a certain range (1 deviation) about the mean. In a normal distribution of players' performance, right about 68% of all data points will fall within +/- 1 deviation of the overall average. (In other words, the vast majority of the players will fall pretty close to the middle while stronger and weaker players work their way to the sides.)
If the data distribution is normal, then the Elo rating system is a pretty good predictor of match outcomes. As players advance their rating, they will eventually begin to match stronger players who are more on their level. Eventually, the probability of a win becomes 50:50 and that ends up being the win:loss ratio.
Within the Builder Base, we see this constantly. The number of trophies you have is effectively your Elo rating. As you win matches, you get more trophies; lose and you drop trophies. You have the ability to chose your pool of opponents by dropping trophies into a lower trophy range, but this still works within the system because you are reducing your rating artificially to go against weaker opponents. Of course, if you simply play the game as designed, you will gain trophies until you reach a sweet spot consistent with the maturity of your game and your skills as a player.
I've spoken of the Four Pillars of Clash before - Offensive Capabilities, Attacking Skills, Defensive Capabilities, and Base Design. The more players match up across these four categories, the closer their trophy rankings will be to one another. While one can compensate to a degree by being stronger in one of the four pillars, that will only get you so far; eventually, someone who is more balanced will beat you and your progress through the trophy rankings will stall or drop.
To provide a more extreme example.... It is possible to go from a brand new game all the way to BH7 in just over a month. Such a player has all the toys, but his defenses will be low - typically L3 or under. His troops will also be limited - he may have one or two troops that he maxes along the way, but that does not provide a versatile attacking position. If this is his only game, he will also be lacking in attacking skills by sheer virtue of not having had much opportunity to attack.
That means he has all the toys of a BH7 but he's lacking in at least two, if not three, of the four pillars. Would you really consider this player BH7? The reality is that he's too weak to be an effective competitor in the upper trophy rankings and will likely be in the same trophy range as a BH4 or BH5 player who is taking the time to better develop his game. I've had plenty of lower level matches where my BH4 or BH5 game comes up against one of these BH7 games and not only wins the match but even gets a 3-star win against the game.
In other words - my lowly BH4 and BH5 games were performing at the same level as a hyper-rushed BH7 player, so we drew what looked like an underdog match on the surface but was actually an even match because we each resided in the same trophy range (had a similar Elo-type rating).
Generally speaking, the differences between two players are not this stark. It tends to be just one level of BH and defenses are a hodge-podge of levels right around the level of the lower BH competitor. The take away here: BH level isn't always a good predictor of the outcome of a match; you're better at the game than your BH level suggests.
The rating system enables you to fight against peers - opponents who perform similarly against other players of the game. Their skills and capabilities, overall, will be a close match to your own. If you find yourself in an underdog match, know that the other players is almost certainly weaker in one of the four pillars where you're stronger and that you DO stand a chance of winning because his/her performance is right about even with your own.
Here's this article's companion video. I'll be back soon with more...!
Comments
Post a Comment