Shoulders

New ranks in rainbow six siege. Ranks in Rainbow Six Siege. How the ranking system works. What affects the change in rank? What parameters are taken into account

The first 5 battles are required to pass calibration, after which you will be assigned a rank.

First, let's look at what ranks there are in Rainbow Six Siege.

As we can see from the table, ranks in Tom Clancy’s Rainbow Six Siege are divided into five types:

  • Copper
  • Bronze
  • Silver
  • Gold
  • Platinum / Diamond

There are a total of 20 ranks in Rainbow Six Siege.

Let's take a closer look at the ranking system and how it works.

How the rank system works in Tom Clancy's Rainbow Six Siege

Rainbow Six Siege's ranking system is based on the TrueSkill algorithm. This is an evolutionary system that has outgrown the Elo/Glico rating. What makes it different from the original systems is that it works better in multiplayer games (Elo and Glico are designed for 1v1 games like chess).

Your skill is initially calculated after 5 games. After this, the rank value is updated after each ranked game.

Players from different regions almost never play against each other, so regionality is relative and ranks cannot be directly compared between regions.

What affects the change in rank? What parameters are taken into account?

Main two things:

  • Player skill levels / overall team level in the game.
  • Did your team win the game.

What doesn't matter:

  • The ratio of deaths to murders does not matter.
  • The score doesn't matter.

Rainbow Six is ​​a team game, so teamwork skills and your overall benefit to the team are taken into account.

In addition, the system takes into account if there is a large difference in skills between the two teams. In this case, at the end of the game, if your team was much superior and dominant, you will not receive a rank. For example, if a team in which players have gold ranks fights with a team from bronze, then if they win they will receive a minimum of points towards their rank, and vice versa. If the bronze team wins, they will receive a lot of points.

There are a few more points to consider:

  • If you leave a ranked game, go AFK, or your team kicks you, you will lose some points, and this will affect your rank.
  • If the game server crashes, and the crash occurs due to an error on the part of the developers, then the ranks will not be affected, and the match is considered not played.

Let's summarize and briefly describe the factors influencing the rank.

Factors influencing the rank of Rainbow Six Siege

  • Player skill levels / Team levels in the game
  • Win or lose the match
  • The ratio of deaths to murders does not matter
  • A large skill difference between teams has a negative outcome based on who won the match.
    • If a team loses to Bronze while in Gold, the second team's rank will change greatly in a negative direction
    • If a team wins a gold team while playing a bronze team, the ranks of both teams will change dramatically.
  • Quitting a game early will always count as a loss towards qualifying, even if your team wins.
  • If you get kicked or AFK, your rank may also be lowered.
  • Connection problems are counted as a negative factor if these problems are the fault of the player and not the developers.

06.12.2016 12:01

As part of Operation Red Crow, we've made some fixes to the rank and matchmaking systems that have raised some concerns among players in the community. We will try to describe in detail the changes in the game since the beginning of the new season.

What is a "skill"?

The skill reflects your chances of winning the game. When comparing this indicator for two teams, it is theoretically possible to determine who will be stronger. The greater the difference, the higher the probability of victory for one of the groups of players. When two teams with the same skill scores meet in a game, their chances are equal. The Rainbow Six Siege algorithm takes into account two parameters: the player's skill score (μ) and deviation (σ). Player skill scores are distributed along a Gaussian curve with an average of 25 (peak at Silver and lowest at Gold).

What is "deviation"?

Player skill rating is an approximate parameter. The more games you play, the more data our team receives about your abilities and the more accurate the assessment will be. The fairness of the calculations is reflected in the value of “sigma” (σ). The lower the deviation, the higher the accuracy.

How are skill and matchmaking rating related?

In Season 4, your matchmaking rating is proportional to your skill (matchmaking rating = 100*μ). Previously, we used a progression system that took into account your involvement in the process when calculating your rank. In other words, if you managed to reduce the deviation value, the rank increased. As a result, when calculating, the differences in skill indicators and current rank could turn out to be significant. For example, a player whose skill is rated as "Silver" actually achieved the rank of "Platinum" by playing a lot of games. As a result, some confusion arose, since games within the same rank were strikingly different from each other, because there was no guarantee that users with the same skills would participate in it. When comparing data from previous seasons and after the implementation of the new system, the discrepancy in the skills of opposing teams on average could be double the size.

Since Rainbow Six Siege is a competitive game, we have made the following decision: Only skill will now be taken into account when assigning matchmaking ratings. Rank itself does not imply any reward, and a player's progress at a certain point only reflects an increase in his gaming capabilities.

How is the matchmaking rating updated after a game?

Only the outcome of the game affects the rating update. If the result is unexpected (for example, you lose to a team of players whose skill is lower than yours), significant changes will be made to the rating, and vice versa: the more confident we are in your rating (if the value of "σ" is low), the less significant the update will be.

Let's look at an example with only two players, Craig and Justin. The following values ​​do not reflect actual updates made by the system, but should only give you an idea of ​​​​its work. Justin received a Gold rank (μ ≈ 29) and maintains this level for quite a long time, and his deviation value is small (σ ≈ 3). Justin takes on Craig (a stubborn Copper player) and loses. Since the deviation is small, most likely Justin was just unlucky, but his skill will still decrease (μ ≈ 28). The value will be this way because the system takes into account the rank of the player to whom you lost. If Justin were defeated by a user with the Gold rank, the changes would be different (μ ≈ 28.5), and if he was defeated by someone with the Diamond rank, the changes would be μ ≈ 28.9.

At the same time, we must not forget that Craig relatively recently joined the community: the system must determine what rating should be assigned to him with indicators μ ≈ 17 and σ ≈ 6. By choosing the Tachankin operative, he was able to win, and his skill indicator will increase to 20. Victory above Justin suggests that Craig is a fairly capable player, and therefore the deviation value should decrease (σ ≈ 5.5). However, two factors contributed to the significant increase in skill: the fact that we are confident in Justin's rating, and the fact that Craig is still a beginner.

Has the game balance been adjusted in the end?

After the launch of the new system, a greater number of balanced games were recorded. In this case, two methods are used for calculation. The first is to determine the average skill of opposing teams. Yes, the threshold for various ranks has been updated, but the value of the skill in the community on average has remained unchanged. Therefore, comparison with the results of previous seasons still allows us to obtain correct data. As a result, after the launch of Operation Red Crow, the situation is developing in a positive direction.

The second method is to determine the balance based on the results of the game. Operation Red Crow is increasingly posting scores of 4-2 or close to it. The most noticeable result is a decrease in the number of games ending in a “dry” victory for one of the teams (by 2.3%). These achievements are proof that, with all the changes in the system for obtaining a new rank, now when selecting players, users with similar skill indicators are increasingly pitted against each other.

How do my results affect the number of points I receive?

No way. If you get 15 kills in several games, but lose every one of them, your chances of winning are generally low, which means your rank will not be high. If a player does not commit a single kill, but wins over and over again, his rating will increase (perhaps he conducts excellent reconnaissance and thereby helps the team neutralize the enemy with ease). The point is that in the long run, it is the ability to play and interact with the group that influences the positive outcome of the game.

Why do my friends and I gain/lose different amounts of points after the game?

If your skill scores are equal, the upgrade will depend on how confident the system is in your position in the ranking. For example, if you both have a Gold IV rank, but your friend has played more games (his rejection rate will be lower), you will gain/lose more points than him.

Why did the threshold for obtaining a rank change?

Since player progress no longer affects matchmaking rating, a value that was easier to calculate and relate to skill was needed. The new rank thresholds have nothing to do with the similar parameters of the third season: comparing them is as pointless as comparing apples with oranges.

Is the number of players with the Diamond rank limited?

No, all players with a Diamond rating are those users who managed to achieve a very high skill score. Based on data from past seasons, they are at the end of the distribution curve. Why does it feel like too many players managed to reach the Diamond rank in Season 4? Because in this case, the progression system no longer has any influence and players do not have to put in a lot of effort to reduce the deviation value.

So, in Season 4 it really became easier to get the Diamond rank?

No, it didn't. It’s just that now the players who deserve it reach their goal faster. To date, 80% of users with the Diamond rank received it last season. Among all participants in ranked games in Season 4, only 0.5% managed to achieve such a high result. Matchmaking rating 3700 IS NOT A WARRANTY the player having the same skill value as a rating of 3700 in Season 3.

Why are ranks reset and how does this affect me?

The previous version of the system had a strong impact on skill assessment. A full reset (resetting the original "μ" and "σ" indicators) allows you to start with a "clean slate" and provide an accurate and correct assessment of your skill.

Moreover, the justification for a complete reset is the significant change in the overall picture within each season.

Will I gain ranks faster when using my alternate account?

Yes, but not immediately after a hard reset. Over time, the variance value for your main account will decrease and we will be able to more accurately determine your skill score. If the playstyles of the main and alternate accounts are similar, then the skill value for the latter will approach the same value for the former.

Does rank count in unranked games? What about the access level?

No. The ranked list of games has its own matchmaking rating based on the same system. Access level does not affect the rating for either rated or unrated games.

Why do I get so few rating points for a 40-minute ranked game?

What is the peak of the distribution? "Bronze"? "Silver"?

The main part is the peak indicator in the “Silver” rank and the lowest in the “Gold” rank:

If I usually play with the same team, how do you determine the difference between my matchmaking rating and that of my allies?

No way. If you play with one team most often, then we will get a fairly accurate estimate of the group's rating. When playing without this team, your personal rating will change, reflecting not skill, but a combination of opportunities to win with and without a team, that is, the likelihood that you will play in a permanent lineup or in a different group. Overall, the changes will reflect your chances of winning.

Did I like the feeling that it took a lot of effort to achieve the Diamond rank? Why was the system changed?

As stated earlier, if your rank is a clearer reflection of your actual skill, the games will be fairer. If a player who deserves the Diamond rank continues to be ranked lower just because he hasn't played as much, it won't be fair to the other Gold rank players.

My rank is “Diamond”, why does the system offer to confront players with the “Gold” rank?

There are not many players with the Diamond rank, and the matchmaking system is not always able to find users with the appropriate skill. So the choice is between how long you are willing to wait for your opponent, and how accurately the balance of the game should be maintained.

We hope that we have explained to you the intricacies of the current system and answered most of the questions that concern the community. After reading, we invite you to join the discussions on the forums and our Reddit page! We will continue to study the situation and in the future we may supplement this text with answers to new questions!

06.12.2016 12:01

As part of Operation Red Crow, we've made some fixes to the rank and matchmaking systems that have raised some concerns among players in the community. We will try to describe in detail the changes in the game since the beginning of the new season.

What is a "skill"?

The skill reflects your chances of winning the game. When comparing this indicator for two teams, it is theoretically possible to determine who will be stronger. The greater the difference, the higher the probability of victory for one of the groups of players. When two teams with the same skill scores meet in a game, their chances are equal. The Rainbow Six Siege algorithm takes into account two parameters: the player's skill score (μ) and deviation (σ). Player skill scores are distributed along a Gaussian curve with an average of 25 (peak at Silver and lowest at Gold).

What is "deviation"?

Player skill rating is an approximate parameter. The more games you play, the more data our team receives about your abilities and the more accurate the assessment will be. The fairness of the calculations is reflected in the value of “sigma” (σ). The lower the deviation, the higher the accuracy.

How are skill and matchmaking rating related?

In Season 4, your matchmaking rating is proportional to your skill (matchmaking rating = 100*μ). Previously, we used a progression system that took into account your involvement in the process when calculating your rank. In other words, if you managed to reduce the deviation value, the rank increased. As a result, when calculating, the differences in skill indicators and current rank could turn out to be significant. For example, a player whose skill is rated as "Silver" actually achieved the rank of "Platinum" by playing a lot of games. As a result, some confusion arose, since games within the same rank were strikingly different from each other, because there was no guarantee that users with the same skills would participate in it. When comparing data from previous seasons and after the implementation of the new system, the discrepancy in the skills of opposing teams on average could be double the size.

Since Rainbow Six Siege is a competitive game, we have made the following decision: Only skill will now be taken into account when assigning matchmaking ratings. Rank itself does not imply any reward, and a player's progress at a certain point only reflects an increase in his gaming capabilities.

How is the matchmaking rating updated after a game?

Only the outcome of the game affects the rating update. If the result is unexpected (for example, you lose to a team of players whose skill is lower than yours), significant changes will be made to the rating, and vice versa: the more confident we are in your rating (if the value of "σ" is low), the less significant the update will be.

Let's look at an example with only two players, Craig and Justin. The following values ​​do not reflect actual updates made by the system, but should only give you an idea of ​​​​its work. Justin received a Gold rank (μ ≈ 29) and maintains this level for quite a long time, and his deviation value is small (σ ≈ 3). Justin takes on Craig (a stubborn Copper player) and loses. Since the deviation is small, most likely Justin was just unlucky, but his skill will still decrease (μ ≈ 28). The value will be this way because the system takes into account the rank of the player to whom you lost. If Justin were defeated by a user with the Gold rank, the changes would be different (μ ≈ 28.5), and if he was defeated by someone with the Diamond rank, the changes would be μ ≈ 28.9.

At the same time, we must not forget that Craig relatively recently joined the community: the system must determine what rating should be assigned to him with indicators μ ≈ 17 and σ ≈ 6. By choosing the Tachankin operative, he was able to win, and his skill indicator will increase to 20. Victory above Justin suggests that Craig is a fairly capable player, and therefore the deviation value should decrease (σ ≈ 5.5). However, two factors contributed to the significant increase in skill: the fact that we are confident in Justin's rating, and the fact that Craig is still a beginner.

Has the game balance been adjusted in the end?

After the launch of the new system, a greater number of balanced games were recorded. In this case, two methods are used for calculation. The first is to determine the average skill of opposing teams. Yes, the threshold for various ranks has been updated, but the value of the skill in the community on average has remained unchanged. Therefore, comparison with the results of previous seasons still allows us to obtain correct data. As a result, after the launch of Operation Red Crow, the situation is developing in a positive direction.

The second method is to determine the balance based on the results of the game. Operation Red Crow is increasingly posting scores of 4-2 or close to it. The most noticeable result is a decrease in the number of games ending in a “dry” victory for one of the teams (by 2.3%). These achievements are proof that, with all the changes in the system for obtaining a new rank, now when selecting players, users with similar skill indicators are increasingly pitted against each other.

How do my results affect the number of points I receive?

No way. If you get 15 kills in several games, but lose every one of them, your chances of winning are generally low, which means your rank will not be high. If a player does not commit a single kill, but wins over and over again, his rating will increase (perhaps he conducts excellent reconnaissance and thereby helps the team neutralize the enemy with ease). The point is that in the long run, it is the ability to play and interact with the group that influences the positive outcome of the game.

Why do my friends and I gain/lose different amounts of points after the game?

If your skill scores are equal, the upgrade will depend on how confident the system is in your position in the ranking. For example, if you both have a Gold IV rank, but your friend has played more games (his rejection rate will be lower), you will gain/lose more points than him.

Why did the threshold for obtaining a rank change?

Since player progress no longer affects matchmaking rating, a value that was easier to calculate and relate to skill was needed. The new rank thresholds have nothing to do with the similar parameters of the third season: comparing them is as pointless as comparing apples with oranges.

Is the number of players with the Diamond rank limited?

No, all players with a Diamond rating are those users who managed to achieve a very high skill score. Based on data from past seasons, they are at the end of the distribution curve. Why does it feel like too many players managed to reach the Diamond rank in Season 4? Because in this case, the progression system no longer has any influence and players do not have to put in a lot of effort to reduce the deviation value.

So, in Season 4 it really became easier to get the Diamond rank?

No, it didn't. It’s just that now the players who deserve it reach their goal faster. To date, 80% of users with the Diamond rank received it last season. Among all participants in ranked games in Season 4, only 0.5% managed to achieve such a high result. Matchmaking rating 3700 IS NOT A WARRANTY the player having the same skill value as a rating of 3700 in Season 3.

Why are ranks reset and how does this affect me?

The previous version of the system had a strong impact on skill assessment. A full reset (resetting the original "μ" and "σ" indicators) allows you to start with a "clean slate" and provide an accurate and correct assessment of your skill.

Moreover, the justification for a complete reset is the significant change in the overall picture within each season.

Will I gain ranks faster when using my alternate account?

Yes, but not immediately after a hard reset. Over time, the variance value for your main account will decrease and we will be able to more accurately determine your skill score. If the playstyles of the main and alternate accounts are similar, then the skill value for the latter will approach the same value for the former.

Does rank count in unranked games? What about the access level?

No. The ranked list of games has its own matchmaking rating based on the same system. Access level does not affect the rating for either rated or unrated games.

Why do I get so few rating points for a 40-minute ranked game?

What is the peak of the distribution? "Bronze"? "Silver"?

The main part is the peak indicator in the “Silver” rank and the lowest in the “Gold” rank:

If I usually play with the same team, how do you determine the difference between my matchmaking rating and that of my allies?

No way. If you play with one team most often, then we will get a fairly accurate estimate of the group's rating. When playing without this team, your personal rating will change, reflecting not skill, but a combination of opportunities to win with and without a team, that is, the likelihood that you will play in a permanent lineup or in a different group. Overall, the changes will reflect your chances of winning.

Did I like the feeling that it took a lot of effort to achieve the Diamond rank? Why was the system changed?

As stated earlier, if your rank is a clearer reflection of your actual skill, the games will be fairer. If a player who deserves the Diamond rank continues to be ranked lower just because he hasn't played as much, it won't be fair to the other Gold rank players.

My rank is “Diamond”, why does the system offer to confront players with the “Gold” rank?

There are not many players with the Diamond rank, and the matchmaking system is not always able to find users with the appropriate skill. So the choice is between how long you are willing to wait for your opponent, and how accurately the balance of the game should be maintained.

We hope that we have explained to you the intricacies of the current system and answered most of the questions that concern the community. After reading, we invite you to join the discussions on the forums and our Reddit page! We will continue to study the situation and in the future we may supplement this text with answers to new questions!

As part of Operation Red Crow, Ubisoft has made some fixes to the ranking and matchmaking systems, which have raised some concerns among players in the community. This article details the changes to the game since the start of the new season.

What is a "Skill"?

The skill reflects your chances of winning the game. When comparing this indicator for two teams, it is theoretically possible to determine who will be stronger. The greater the difference, the higher the probability of victory for one of the groups of players. When two teams with the same skill scores meet in a game, their chances are equal. The Rainbow Six Siege algorithm takes into account two parameters: the player's skill score (μ) and deviation (σ). Player skill scores are distributed along a Gaussian curve with an average of 25 (peak at Silver and lowest at Gold).

What is "DEVIATION"?
Player skill rating is an approximate parameter. The more games you play, the more data our team receives about your abilities and the more accurate the assessment will be. The fairness of the calculations is reflected in the value of “sigma” (σ). The lower the deviation, the higher the accuracy.

How are skill and matchmaking rating related?
In Season 4, your matchmaking rating is proportional to your skill (matchmaking rating = 100*μ). Previously, we used a progression system that took into account your involvement in the process when calculating your rank. In other words, if you managed to reduce the deviation value, the rank increased. As a result, when calculating, the differences in skill indicators and current rank could turn out to be significant. For example, a player whose skill is rated as "Silver" actually achieved the rank of "Platinum" by playing a lot of games. As a result, some confusion arose, since games within the same rank were strikingly different from each other, because there was no guarantee that users with the same skills would participate in it. When comparing data from previous seasons and after the implementation of the new system, the discrepancy in the skills of opposing teams on average could be double the size.
Since Rainbow Six Siege is a competitive game, we have made the following decision: Only skill will now be taken into account when assigning matchmaking ratings. Rank itself does not imply any reward, and a player's progress at a certain point only reflects an increase in his gaming capabilities.

How is the matchmaking rating updated after a game?
Only the outcome of the game affects the rating update. If the result is unexpected (for example, you lose to a team of players whose skill is lower than yours), significant changes will be made to the rating, and vice versa: the more confident we are in your rating (if the value of "σ" is low), the less significant the update will be.
Let's look at an example with only two players, Craig and Justin. The following values ​​do not reflect actual updates made by the system, but are intended to give you an idea of ​​how the system is performing. Justin received a Gold rank (μ ≈ 29) and maintains this level for quite a long time, and his deviation value is small (σ ≈ 3). Justin takes on Craig (a stubborn Copper player) and loses. Since the deviation is small, most likely Justin was just unlucky, but his skill will still decrease (μ ≈ 28). The value will be this way because the system takes into account the rank of the player to whom you lost. If Justin were defeated by a user with the Gold rank, the changes would be different (μ ≈ 28.5), and if he was defeated by someone with the Diamond rank, the changes would be μ ≈ 28.9.

At the same time, we must not forget that Craig relatively recently joined the community: the system must determine what rating should be assigned to him with indicators μ ≈ 17 and σ ≈ 6. By choosing the Tachankin operative, he was able to win, and his skill indicator will increase to 20. Victory above Justin suggests that Craig is a fairly capable player, and therefore the deviation value should decrease (σ ≈ 5.5). However, two factors contributed to the significant increase in skill: the fact that we are confident in Justin's rating, and the fact that Craig is still a beginner.

Has the game balance been adjusted in the end?
After the launch of the new system, a greater number of balanced games were recorded. In this case, two methods are used for calculation. The first is to determine the average skill of opposing teams. Yes, the threshold for various ranks has been updated, but the value of the skill in the community on average has remained unchanged. Therefore, comparison with the results of previous seasons still allows us to obtain correct data. As a result, after the launch of Operation Red Crow, the situation is developing in a positive direction.

The second method is to determine the balance based on the results of the game. Operation Red Crow is increasingly posting scores of 4-2 or close to it. The most noticeable result is a decrease in the number of games ending in a “dry” victory for one of the teams (by 2.3%). These achievements are proof that, with all the changes in the system for obtaining a new rank, now when selecting players, users with similar skill indicators are increasingly pitted against each other.

How do my results affect the number of points I receive?
No way. If you get 15 kills in several games, but lose every one of them, your chances of winning are generally low, which means your rank will not be high. If a player does not commit a single kill, but wins over and over again, his rating will increase (perhaps he is excellent at reconnaissance and thereby helps the team neutralize the enemy with ease). The point is that in the long run, it is the ability to play and interact with the group that influences the positive outcome of the game.

Why do my friends and I gain/lose different amounts of points after the game?
If your skill scores are equal, the upgrade will depend on how confident the system is in your position in the ranking. For example, if you both have a Gold IV rank, but your friend has played more games (his rejection rate will be lower), you will gain/lose more points than him.

Why did the rank threshold change?
Since player progress no longer affects matchmaking rating, a value that was easier to calculate and relate to skill was needed. The new rank thresholds have nothing to do with the similar parameters of the third season: comparing them is as pointless as comparing apples with oranges.

Is the number of players with the DIAMOND rank limited?
No, all players with a Diamond rating are those users who managed to achieve a very high skill score. Based on data from past seasons, they are at the end of the distribution curve. Why does it feel like too many players managed to reach the Diamond rank in Season 4? Because in this case, the progression system no longer has any influence and players do not have to put in a lot of effort to reduce the deviation value.

So, in Season 4 it became easier to get the DIAMOND rank?
No, it didn't. It’s just that now the players who deserve it reach their goal faster. To date, 80% of users with the Diamond rank received it last season. Among all participants in ranked games in Season 4, only 0.5% managed to achieve such a high result. A matchmaking rating of 3700 DOES NOT guarantee that a player will have the same skill value as a 3700 rating in Season 3.

Why are ranks reset and what effect does this procedure have on me?
The previous version of the system had a strong impact on skill assessment. A full reset (resetting the original "μ" and "σ" indicators) allows you to start with a "clean slate" and provide an accurate and correct assessment of your skill.
Moreover, the justification for a complete reset is the significant change in the overall picture within each season.

Will I gain ranks faster when using my alternate account?
Yes, but not immediately after a hard reset. Over time, the variance value for your main account will decrease and we will be able to more accurately determine your skill score. If the playstyles of the main and alternate accounts are similar, then the skill value for the latter will approach the same value for the former.
Thanks to the new matchmaking ranking system, the current rank on an alternate account will be achieved faster, reducing the negative impact of having such accounts overall.

Does rank count in unranked games? What about the access level?
No. The ranked list of games has its own matchmaking rating based on the same system. Access level does not affect the rating for either rated or unrated games.

Why do I get so few rating points for a 40-minute ranked game?
The matchmaking rating system does not imply any reward. If we do not receive new data to evaluate your skill, then the changes will be minor.

What is the peak of the distribution? "Bronze"? "Silver"?
The main part is the peak indicator in the “Silver” rank and the lowest in the “Gold” rank:

If I usually play with the same team, how do you determine the difference between my matchmaking rating and that of my allies?
No way. If you play with one team most often, then we will get a fairly accurate estimate of the group's rating. When playing without this team, your personal rating will change, reflecting not skill, but a combination of opportunities to win with and without a team, that is, the likelihood that you will play in a permanent lineup or in a different group. Overall, the changes will reflect your chances of winning.

I liked the feeling that I had to spend a lot of effort to achieve a rank. Why was the system changed?
As stated earlier, if your rank is a clearer reflection of your actual skill, the games will be fairer. If a player who deserves the Diamond rank continues to be ranked lower just because he hasn't played as much, it won't be fair to the other Gold rank players.

My rank is “DIAMOND”, why does the system offer to oppose players with the rank “GOLD”?
There are not many players with the Diamond rank, and the matchmaking system is not always able to find users with the appropriate skill. So the choice is between how long you are willing to wait for your opponent, and how accurately the balance of the game should be maintained.

We hope that we have explained to you the intricacies of the current system and answered most of the questions that concern the community.

GameGuru

You can only get the full portion of impressions and emotions that Rainbow Six: Siege can give to a player in the company of friends. Individuals in most situations are doomed to failure: even if there is voice communication in the game, the degree of synergy in the team necessary for victory can be achieved only if there is complete and mutual trust on the part of all its participants. But a “stack” of experienced fighters under the command of a good tactician works real miracles...

Read the full review

gambling addiction

There are reasons to be sad. Siege is too unusual, very thoughtful, costs four times more than Counter-Strike: Global Offensive, and even then it has microtransactions that do not make you stronger (skins, accelerating the flow of points to unlock operatives), but purely psychologically cause rejection. And yet in multiplayer, this Rainbow Six is ​​perceived as it should be. If only there was someone to play with.

Read the full review

Strategy

In their project “Tom Clancy's Rainbow Six: Siege”, Ubisoft studio still managed to implement its plans quite well. This is truly a new tactical team action movie, which has no worthy analogues today. And the word team here can be highlighted in double bold font, so How the game will give you a lot of incomparable sensations when performing operations with your comrades, however, it can be quite disappointing when playing alone.

Read the full review

Gmbox

By releasing Siege, Ubisoft succeeded in the main thing: making a game that found its own niche in the online shooter market. This is not high-speed Counter-Strike or Call of Duty, not a large-scale Battlefield and not some smart entertainment like Insurgency or ArmA, but something else. Very familiar and understandable from the first seconds, but at the same time fresh. The content that is already in the game is good: the game has great maps, unusual fighters and fresh mechanics. Unfortunately, this is still too little: as in the case of the recent Star Wars Battlefront...

Read the full review

IGN Russia

Once you've unlocked all the operatives and explored the maps, you'll likely find no reason to return to the game. Rainbow Six: Siege is perfect for a group of friends who, as children, hid behind their beds and built barricades out of pillows, imagining themselves as special forces on top-secret operations. In this game you will feel like a real complementary team and will be able to play out a variety of offensive and defensive tactics.

Read the full review

[email protected]

The main problem of Tom Clancy's Rainbow Six: Siege will probably be solved by itself in a few months, when the “casual” players leave, and an experienced and capable community will be formed from the remaining ones. In the meantime, it’s better not to try Rainbow Six: Siege without a team of loyal comrades who are ready have your back at any time.

Read the full review

3DNews

The tragedy of Siege is not in the cards and possibilities, but in the monotony. It doesn’t matter what task the special forces are tasked with: defusing a bomb, saving a hostage, or eliminating everyone - all of these are, in fact, small variations of the same mode. Almost always, what happens comes down to the complete destruction of opponents. Why take out a civilian if it’s easier to kill everyone? Of course, “Anti-Terror” is available in the menu...

Read the full review

GameMAG

We have already seen several examples where large publishers tried to sell an F2P game under the guise of a full-fledged one. With all due respect, Evolved, Titanfall, PvZ, Battlefront can hardly be called complete products that deserve their price tag. But unlike them, Siege has a much better chance of becoming a long-playing record that's fun to play with friends months later. However, the new Rainbow Six still does not reach the level of a full-fledged AAA project.