Friday, April 24, 2020

Multiplayer Competitive Matchmaking!

After a lot of hard work, and great support and feedback from the community of Cribbage Pro players during "beta testing", we are excited to announce the full release of Online Multiplayer Competitive Matchmaking in Cribbage Pro!


The game of cribbage has often faced challenges in calculating player skill levels in both real world physical games and online digital versions. Accurately ranking players by their skill is a challenge in any game that involves a degree of chance/luck. We have tried many different ways of doing this in the past, and through that experience we have learned a lot. With that history as a teacher, and algorithms now available that have improved dramatically since we started, we are happy to announce our new Competitive Rank - CR (or "Cribbage Rank" if you prefer) ranking system.

Here are the new basic elements of Competitive Matchmaking:

  • Best-Of-3
  • Automatic Counting
  • 10 Sec. Turn Limit
  • No Chat
  • Start Automatically
  • Disconnect/Quit is Always a Loss
  • Classic (non-competitive) XP Level of 5+ Required
  • 5 Placement Matches Required for Ranking

We chose these criteria to make competitive matchmaking the best it can be. First, it is a matchmaking system so that it avoids players selecting only newer players to try and easily win points against, while also providing fun and challenging games. We went with a "Best-of-3" series as a good way to remove some of the chance/luck from the game (alternating dealer as expected). There are also other things our algorithm does internally to help offset the chance/luck factor. Automatic Counting was chosen because, although counting and math is certainly a skill, the skill we want to measure here is playing cribbage. We have found that although new players may miss points from time to time, a skilled player will win against an unskilled player in a best of 3 series no matter if manual counting and muggins is enabled or not. The automatic counting option enables us to set a shorter time limit of 10 seconds, which makes a best of three games series more practical. The shorter time limit helps move the game along so that the entire match can still complete in a reasonable time frame. We removed the chat for the same reason - to keep it fast and competitive without unnecessary distractions.

This new Competitive Matchmaking is also a "Season" based system. We will be having "Competitive Seasons", just like you would see in a professional sports season. This means that on a regular basis (currently about 3 months; Quarterly), a new season will start and give everyone a chance to come in and participate on equal footing to see if they can take the title of the best player of the season! We are also planning on awarding a special prize to the top finishers of the season! The specifics of that prize will be shared at a later time and may be unique for each season. You must play games every week of the season (2 per week minimum), or you will see a reduction in your CR as a result. This happens because of an increase in our uncertainty of your actual skill if you don't keep playing regularly against other players. Stale players will not be able to sit at the top. Make certain you are regularly playing competitive matches to hold on to your CR.

When you complete your placement matches, you will be given a CR value, and a corresponding named CR Level. The levels are simply a breakdown of the CR value like this:

Bronze - 1 - 1499 CR
Silver - 1500 - 1999 CR
Gold - 2000 - 2499 CR
Platinum - 2500 - 2999 CR
Diamond - 3000 - 3499 CR
Master - 3500 - 3999 CR
Grand-master - 4000+ CR

The system that is used to calculate your CR, and to adjust it when you win or lose (and as part of the matchmaking itself), is similar to the "Elo Rating System" but a bit more complex like a derivative of that called the "Glicko Rating System" (to clarify, it is NOT just the Glicko system). We have taken the ideas of those systems, and others that are derived from them in modern games/sports, and created our own customized system for cribbage games. Instead of repeating what you can read in each of those links to those systems, just know that we have a system that is very much like them that includes a "confidence factor" (or sometimes called an "uncertainty factor") and a few other adjustments specific to what we felt was needed after running various simulations/tests of the system. We will almost certainly make further tweaks and adjustments to those things as we review how the system is working from time to time.

There are some questions that are common with a system like this, so I'll clarify a few of them here. First, you will see situations where someone's win percentage may be lower than you (and/or have fewer games played than you), and yet they are ranked higher. This is because a skill ranking system like this is significantly concerned with the skill of the opponent you are facing when determining how many points you win or lose from each match you play. Someone who has faced many highly skilled players, and won, will not have to play as many games to move up. Similarly, if they are relatively lower skilled and play higher skilled, they will lose very little when they lose games. This system can then account for people who play a lot of games, versus someone playing fewer total. As long as you are playing games every week, that is enough for the system to calculate your skill with a high enough degree of confidence (just don't miss too many days). You will also notice that in your stats there will not be any "skunk" related stats for competitive matchmaking games, as currently all "best of 3+" series style games have no concept of a skunk in them. You always win or lose by the same minimum games in a series based system like this, for instance a best-of-3 is always won by whoever wins 2 games, a best-of-5 by whoever wins 3, etc. You can never win a a best-of-3 series by more than 2 games, or less than 2 games, so as such there is no way to "skunk" someone in a series.

Lastly, a disconnect or forfeit in a competitive game, of any kind, will count as a full loss in every respect. There is no mercy here in competitive play. Consider this like a professional sports league, and there will be no exceptions beyond the rare case where we can accurately determine that there was a bug of some kind in our system that caused the disconnect/forfeit (which you must fully report to us for investigation). To be clear, this does NOT mean that you will avoid a loss if you try to show it was your internet connection failing - that will still be a loss as there would be too many ways to cheat the system otherwise. We will determine independently if it was our code itself that caused the problem, and only then will a loss be reversed if needed.

This new Competitive Rank system continues to extend our goal of making Cribbage Pro the best option for serious cribbage players. We hope you all enjoy this new mode of play as much as we have during our testing of it. You should find the matches to be against similarly skilled players, so that the matches are both challenging and fun! As always, if you have any questions or suggestions, please let us know through email to support@FullerSystems.com

101 comments:

  1. I would like to see 15 second timer. 10 seconds just seems a little too fast and feel like im playing speed cribbage with not a lot of room for thinking on those tough hands one gets every now and then and also having to consider board position, etc.

    ReplyDelete
    Replies
    1. For me, the 10 sec is OK during play but I'd like to see 15secs for the initial card selection..

      Delete
    2. If you're not ready for the 10 sec turn then you're not good enough to play in this league. There's people that as a way of retribution take the whole 9 secs as a way of getting back at you.
      IMO

      Delete
    3. Hi Dads - methinks you are a better player than I. How do you make proper decisions in such a short amount of time? what if you change your mind? Do you not go "overtime" Jerry

      Delete
  2. Would preferr if it would be reduced to 5 sec

    ReplyDelete
  3. I like the 10 sec rule. But what is the penalty for going over? I’ve played at least two games in competitive matchmaking where my opponent went slightly over.

    ReplyDelete
    Replies
    1. You will see a warning the first time you go over, and if after the second warning of the game you still don't play, you will forfeit the game as a loss - just like regular games.

      Delete
    2. Are you saying there is no penalty for playing 15 seconds per hand for example

      Delete
    3. There is absolutely a penalty for playing too long for each hand. You can only go over 10 seconds twice for the entire match. Anything beyond that and you will be forced out and forfeit the game.

      Delete
  4. I have been kicked out of 4 games, randomly. No internet problem. The game gave me losses, and locked me out for 5 minutes. It is a bug that makes the ranking inaccurate

    ReplyDelete
    Replies
    1. Hi Greg, I believe you also wrote a review mentioning the same problem. Please email us at support@FullerSystems.com so we can investigate this further with you.

      Delete
  5. I haven't been able to play for some time on a regular basis on your normal cribbage pro games because of poor connections - is this an additional game and completely separate prom the normal game?

    ReplyDelete
    Replies
    1. Hi, this is inside the same game/app. If you are having connection problems in the game, and your internet is generally good otherwise, please email us at support@FullerSystems.com so we can try to investigate your issue and see if we can offer some suggestions to get connected and playing again.

      Delete
    2. I agree 15 seconds at least for the discard phase

      Delete
    3. I also welcome 15 seconds on the discard phase.

      Delete
  6. Ten seconds is too short. I have played many many games with the 15 limit and don't ever recall feeling my opponent played too slow

    ReplyDelete
    Replies
    1. I agree. 15 seconds would be best. It's rare to use that much time but would be beneficial in those occurrences

      Delete
  7. 10 sec time limit is perfect! Love the pace!

    ReplyDelete
  8. Just a thought: you could have a bracketed tournament at the end.. sort of a Cribbage March Madness

    ReplyDelete
    Replies
    1. 10 seconds can be too short in man mug on occasion but with man. counting, even with a hand that needs a second look, is fair.

      Delete
  9. Are you guys sure the systems algorithm is on ? The miraculous cuts for my opponent get a little old

    ReplyDelete
    Replies
    1. There is no algorithm for shuffling, it is just a random shuffle. Nothing has changed with the fairness of the game in any way. Always true random shuffling and nothing more.

      Delete
    2. Sorry, after playing over 100 matches on this new version, there is no way you can convince me that there isn't some kind of algorithm. The miraculous cuts for any opponent I play lead me to believe that, for some bizzare reason, there is preference of some kind.

      Delete
    3. I agree with this 100%

      Obtw, did you know your connection to this blog is not secure?
      This will be my last post until you fix your issue.

      Delete
    4. @Dad, although HTTPS certainly isn't needed for a blog like this, it has been available for a while so I have gone ahead and turned it on by default. In any case, never share any personal, private or other similarly protected information on our blog (or really any other).

      Delete
    5. Ok thanks for that Mom. SMH

      Delete
    6. I'd say that over 50% of the games I play the miraculous cuts my opponents is amazing towards the end of the game. It's getting to point of not wanting to play on your platform anymore. I've been playing for a long time, real and this app. I'm not making this crap up.

      Delete
    7. To clarify, the response with "@Dad" was simply a mention of the profile name you are using in the comments so that it is clear who we are replying to.

      We have worked very hard to keep the game fair and honest in every respect. Not only have we discussed this in detail in the help/FAQ and multiple blog posts, we have also audited the game (and published those results publicly) to show it is fair while also offering anyone the chance the view the very code itself used in the game. If there is something more you can think of that would help convince you that the game is being fair, just email us your questions at support@FullerSystems.com We are fully committed to the fairness and integrity of the game.

      Delete
    8. You know who I am.
      When I started playing CMM I was in the low 500's, now I'm approaching 1100. I've been playing for over 40 years, just my experience and knowledge should keep me lower than I am right now. But these amazing hands and ridiculous cuts are getting annoying. There's no way I should get 6 hands in a row of 2, 3, or 4 points. I'm not the only one too. If your system is so fair, why are there so many people complaining about it? If your system was as fair as you say it is, no one would be complaining. Just my opinion.

      Delete
    9. @Dad, can't say I know who you are, but hopefully you can see from my responses that we are committed to the integrity of the game. I understand you feel something is off and that others do too. There are others who disagree. When we have two sides disagree on something like this, how do we decide who is right? The only thing we can do is analyze the data to see. If there is a bias, the data will/must show it. This is why we have shared so much. If you can think of other data analysis that would help, we are always open to investigate. The reason we do the audits and investigations is to make certain it is being random and fair at all times.

      Delete
  10. I just started. And my screen went blank.
    Not worth it.

    ReplyDelete
    Replies
    1. Hi, not sure why your screen would just go blank, but please email us at support@FullerSystems.com so we can work with you to get this fixed.

      Delete
  11. Very good improvement.

    Two suggestions and one bug:
    1. Add option to exit/forefit a game you're way behind. If you're up 1-0 and behind by 30 you both may want to move to the third game.
    2. Add indication, perhaps on the cribbage board of the match score (0-0, 1-0, 0-1, 1-1)
    3. System still gets hanged once in a while at 'waiting for oponent' after match is found. Only way out is to forefit and be locked for 5min.

    ReplyDelete
    Replies
    1. Hi, thanks for the suggestions. It is something we will consider in the future. For the match score, that should be shown on the board with game pegs as this is a best of 3 series. If you are not seeing that, please send us a screenshot so we can investigate it further. Similarly, if you ever see it "hang" on "waiting for opponent", please also email us so we can investigate the circumstances more fully. It could just be a connection problem, but if there is a bug of some kind we want to find it and squash it as quickly as we can.

      Delete
    2. If you look at the start line, the pegs to the left inidate game wins 0-1 1-1 etc

      Delete
  12. Your game is buckshit. C'mon play fair. I hate computerized games. You'll never know who's cheating. Not this anymore..

    ReplyDelete
    Replies
    1. Hi, we have done extensive work to keep the game fair and honest in every respect, and it is not cheating or manipulating anything in any way. You can read about all that we have done and the multiple audits we have performed on the system that make sure it is always being fair. If you read all that we have shared and still have questions, please feel free to email us to ask for anything at support@FullerSystems.com

      Delete
  13. Great job getting this set up! My only complaint is that I now realize how bad I am comparatively at this game, lol

    ReplyDelete
  14. Hi,
    Regarding playes who disconnect in CMM. The penelty may not be significant enough. Especially players who disconnect on the last hand when their loss is certain. How about 3-strike rule for disconnects when oponnet is at 100 or above on last game? If done 3 time get them out of CMM leauge.

    ReplyDelete
    Replies
    1. Hi, the penalty is simply that they lose and lose the full rank points they would if played to the end. Anyone disconnecting is certainly not getting ahead by doing so. You are simply getting an "easy win" when that happens.

      Delete
    2. what's wrong with disconnecting? I was at 65-117 and disconnected at the start of the deal because there's no possible point in playing on and wasting everyone's time when skunks aren't even counted. Despite this, I was in fact subject to a 5 minute disconnect penalty.

      there really should be a resign button as unless you keep track of skunks, there is no point in playing on.

      Delete
  15. Agree with the resign suggestion. The problem with disconnecting is that the other guy has to wait about 1minute to timeout.

    ReplyDelete
    Replies
    1. You can resign a game today. On iOS that is the "quit" option in the top left, and on Android you just use the back button. This will prompt you to confirm the forfeit. You are correct that disconnecting will cause a wait for reconnect.

      Delete
  16. what's the logic on 'competitive matches'? I have 5 consecutive matches against lower-ranked players, 4 of them 400+ lower. Is there some kind of logic whereby if two players are closely ranked then it will match quickly, but as time goes on you're more likely to get a much lower ranked player? Or does it just match against the first 'compatible' player?

    ReplyDelete
    Replies
    1. Hi, the matchmaking system is fairly complex. First, it will try to find the best match to provide a good competitive challenge. The criteria for that are fairly open for your first few placement matches. Then, as it waits to find a match, it will loosen that criteria in various ways to eventually find an acceptable match. Generally speaking, the longer the wait the lower the match quality may be, but the skill ranking system will account for that and adjust your rank appropriately based on the expected outcome.

      Delete
  17. If you can't figure out your hand and what to throw in the crib in 10 seconds then maybe stick to playing heads up. Not competitive.

    ReplyDelete
  18. Great cribbage app and nice addition with the CMM. A few question and comments:
    1 - I like the 10sec clock, although there are those occasional hands 2-3 extra seconds would help. Even though I've gone over 10sec to decide on crib discard, I've never lost a game because of using too many time freebies.
    2 - Does the algorithm for choosing a pone take into account whether a particular player was played before? Does it try to avoid repeat matches between players?
    3 - I like stats, so thanks for adding the streak data.
    4 - I would prefer to see all games played in CMM added with the Classic into a combine live play stat box with skunks included. And I would like to see the computer stats NOT included with live play stats. Playing the computer is very different than a real person so the stats would reflect this and therefore shouldn't be combined.
    Thanks for continually keeping it interesting.

    ReplyDelete
    Replies
    1. Hi, since you numbered your comments, I'll reply in the same way.

      1.) I agree. This is exactly why the extra "mercy" time exists.
      2.) Yes, it tries to avoid matching you with someone you played "recently". That is an algorithm and not a simple number, so just know that in general it strongly prefers finding you a new person to play against.
      3.) You're welcome!
      4.) Computer play is definitely different, and why we break things out for each mode of play. You could of course add things up yourself, but I appreciate your suggestion to provide some kind of "total for live games" option. I'll add it to our list of things to consider in the future.

      Delete
  19. Love the game. For competitive match rank, it would be nice to know the number of total players. For example, a rank of 50 out of 2000 is better than a rank of 50 out of 100. Keep up the good work.

    ReplyDelete
  20. Replies
    1. I'm sorry to hear that you didn't enjoy the game for some reason. If you have any suggestions for things you would like to see improved or changed in some way, just email us at support@FullerSystems.com We are always working to improve the game for everyone!

      Delete
  21. I just found this app and I love it.
    I just need to know how to make the
    chat board pop up?

    ReplyDelete
    Replies
    1. Hi, in a multiplayer game with chat enabled, there will be a small yellow chat icon at the top near your opponent's avatar image. Select the yellow chat icon to open the chat window.

      Delete
  22. Hello. Love the app. I am confused about the rankings in the new Competitive Matchmaking. I am currently 78-63 with a CR of 3464. The current leader has a CR of 4301. Is it very unrealistic to catch him at this point? I have a hard time with players that have similar or even worse records than me and are in the top 50 overall. I am no where near their CR level and feel it's impossible to catch them at this point. How does one get their CR level up quicker? Thanks

    ReplyDelete
    Replies
    1. Hi, as mentioned to others with a similar question, this is skill ranking system so it is not based on a win percentage or anything else like that. The only way you increase CR is by winning against someone else, and it will go up more when playing against someone higher than you. The matchmaking system will focus primarily on finding someone relatively close to your rank so it is a challenging match for both players. The point here is that not everyone can "catch" someone else, as it is not a simple matter of playing more games/etc. You must actually win more against the higher ranked players, and so I can't provide any particular strategy other than the general strategy tips we provide in the Help/FAQ and essentially just "win more". It is not unrealistic that someone else will rise to the top of the leaderboard before the end of the season, but we will have to wait to see. Remember, once the new season starts, we are all back at equal levels.

      Delete
    2. This comment has been removed by the author.

      Delete
  23. Wonder if you have any thoughts about the leaderboard? Particularly the top player has 37/9, which is about 50,000-1 by random chance at 50/50 win chance.
    So with 50,000 random players you'd likely get a result like that.

    The fourth player has won 212 of 328, which is 30 million to 1 at the same odds. But their CR is 150 points less.

    It seems that at this point if the leader was concerned about winning, then if they only play the minimum number of games till the end of the season, then that would be inevitable, yet they have done much less to prove their superiority than the lower players.

    The problem being that Cribbage is at least as much luck as skill, and especially with the top players making it best of 3 doesn't really change that - if one player has a 2% edge then going to best of 3 rather than 1-0 increases their chance of winning only from 52 to 53%

    ReplyDelete
    Replies
    1. In short, yes we are watching the leaderboard for this first season and will make adjustments as we see necessary for future seasons. Our analysis shows the skill/luck factor to be about 60/40 in cribbage, and although you are correct that making it a "best of 3" is only helping a little, it does help some and going with a best of 5 or 7 would be too burdensome. The other thing to keep in mind is that because this is a matchmaking system as well, it is forcing higher ranked players to play other higher ranked players to help determine if the ranking is "valid" while including the all important confidence factor in that ranking. This means that the highest ranked player can't maintain that position unless they regularly play and win against other high ranked players. If you have some specific suggestions you think might help with any of your concerns, please let us know as I'm sure we will be tweaking things as the seasons progress.

      Delete
  24. IM NO DOUBT THEE BEST CRIBBAGE PLAYA EVER TO HAVE PLAYED AS WELL AS THE FASTEST

    ReplyDelete
  25. What does buying gold do for you

    ReplyDelete
    Replies
    1. Hi, Cribbage Pro Gold is what is used in the Contests system where you can play for real cash prizes. It is the "Contests" tab/section of the online multiplayer.

      Delete
  26. Is there a way to add the reshuffle button to the competitive cribbage? ACC requires that the deck be shuffled at least three times, and the pone has the right to shuffle. (ACC Rules 2.1-3)

    ReplyDelete
    Replies
    1. The shuffle done here is a full random shuffle each and every time, so although the rules from the ACC make sense when applied to physical card games, it really doesn't apply here. Shuffling multiple times would always yield the same result as the first shuffle - a fully randomly shuffled deck. We added the "Reshuffle" to single player a while back because it was fairly trivial to do and gives some people a sense of control, but even there it really doesn't do anything material - you always have a fully randomly shuffled deck. We don't want to add this to the multiplayer game because it would add several extra waiting points where there would need to be this needless exchange between the dealer and pone about shuffling the deck, if the pone wanted another shuffle, etc. Since such a task is absolutely pointless in a truly randomly shuffled deck, we don't include it and have no plans to do so. Adding pointless things that have no positive impact on the game but instead make the game more frustrating and take longer is just not something we think anyone really wants.

      Delete
  27. hi, further impressions are that after a while it seems to me that the players ranked, say 4000+ seem to be a lot better than those ranked 3900, while those ranked 3700 don't seem to be much different from those at 3900.

    I think CR should be capped somehow for players who have played few games. Perhaps show a 'provisional' CR to such players, but make it clear that they won't get the full CR unless they play more games. Probably the number here should increase over time, so you wouldn't cap CR for players with fewer than 100 games after, say, 1 week of play, but you would after a month. Players who reach the full 'season' cap wouldn't need to play any more games, because statistically they would have proved themselves.

    Looking at the current leaderboard:

    #2 - 57 games (and for a long time number 1)
    #7 - 95 games
    #9 - 95 games
    #13 - 92 games
    #16 - 40 games
    #17 - 78 games
    #22 - 55 games
    #42 - 76 games
    #43 - 36 games
    #47 - 64 games
    #49 - 83 games

    Maybe cap all these players CR at 3999?

    Player #43 doesn't seem to be a particularly great player, even, with a losing record vs Brutal AI (best players win 55%), and 51% only in classic multiplayer.

    ReplyDelete
    Replies
    1. I know we have said this before, but I'll mention it again in the context you provided. This is a ranking system and not a points system, or a system based on win ratio. It does not make sense to look at those things and judge whether someone is "better" than someone else based on those things. The thing to use is the CR, as that is why the system was created. Someone could definitely have a total losing record and still be ranked higher because of who they won against relative to who they lost against. Your suggestion about a "provisional" CR is in place when you are required to play your first 5 placement matches to be ranked. Through our own analysis, we feel that 5 matches is a good balance of the amount of time required to start ranking and the need to have some general idea of where to place someone. When it comes to matchmaking, we absolutely do use a "cap" (and a minimum), however it just doesn't make sense to artificially cap a CR in a skill based ranking system. The algorithm itself will impose the cap/maximum. Please understand, I think it would be great if we could get everyone to play more games, as the more they play the more confidence the system will have in the ranking, but there are other things we are looking at to help improve that over time and in future seasons.

      Delete
    2. Hi, I'm fully aware that the matchmaking win % is not particularly important. I was referring, however, to the win % vs. the single-player AI, which is a player of constant skill against all human opponents.

      I agree that the 5 placement matches will give a good measure of approximately where someone should be, but for example while I definitely feel I am NOT the best player in the game (ranked 50-75 typically), I have runs of 5 or 6 wins against better players (CR 3900+), and those runs are mathematically inevitable (if you toss a coin 100 times, you won't possibly get HTHTHTHTHTHTHTHTHTHTHTHT inevitably, and you might end up with 60 heads and 40 tails, for example).

      The standard error in general increases with the square root of the number of games played, so it follows that in general we can be twice as certain about a player who has played 400 games compared to one who has played 100. And a player who has played just 36 games is much more likely to be experiencing good luck.

      I wonder if, btw, you are catering for the bias towards the player who wins the deal? If in general (per your own data), two players with the same skill level win 56% of games in which they deal, it follows that at best-of-3, the player who won the cut has a 53% of chance of winning the game. Clearly winning is good, and should increase your CR. But it seems perhaps there should be a small extra weight for overcoming the loss of the toss?

      Perhaps it evens out in the long run, but again this comes back to the issue I raise above about uncertainty with players who have not played many games

      Delete
    3. We try to never assume anything about a players single player performance (or "classic" game performance) when it comes to this competitive matchmaking system as people might be playing in a much more casual way there (not caring so much about the outcome). There are also quite a few players who only played a few single player games when they first started and simply don't even play single player at all now. Perhaps we will adjust this in the future, but for now we feel it is fair to only consider the competitive matches themselves.

      You are correct that runs of "luck" happen, and we can only do so much to account for that in the standard rule set of cribbage. To your point, playing more games is probably the best way to do that, and that is true in both the broader sense as well as in the single match case where the dealer has advantage. It would be great if we could run the competitive matchmaking in a longer set of games per match which would help too, but that just isn't going to be practical. Instead, we want to incentivise playing enough games during the season to get a "reasonable" confidence of ranking. Is this currently perfect? Certainly not. Perhaps adding some sort of "bonus" for winning even when under the disadvantage is something we can consider, but we don't want to be adjusting too many things between the seasons or we will lose track of what impacts the changes are having, and we have a few things we are planning to modify for next season already. If you have some other suggestions, we definitely want to hear them. Please email us at support@FullerSystems.com

      Delete
    4. I wasn't suggesting to take into account single player performance, just saying that certain players with high ranking and only a few games played AND a relatively weak performance in other modes, we might conclude that they are most likely currently lucky, rather than one of the best players overall.

      By the way,is it possible to show a percentile or similar? If you are ranked say 200, then is this the top 10%, 5%, 20%, or what?

      Delete
    5. Yes, I understand you are not necessarily asking that we use their single player points (or "classic" points) in competitive matchmaking, but then at the same time you are saying that if you do use it as a factor to know their skill then it seems they may be currently lucky. The point is that you shouldn't ever use the single player stats to indicate multiplayer performance, as it just isn't being fair without knowing the circumstances of each individual. You may be right, and they may be currently just lucky, but I don't think that is a fair way to prove it to be the case. The only way to really know would be to study each game in detail to see the exact cards each had and how they played them. I suppose this is just the technical side of me shining through in this, and frankly it doesn't really matter. Your point is still well taken, in that we would like to find more ways to limit the "luck" factor more often to get to the best ranking possible.

      For percentile in the leaderboard, that is a great idea. I will add it to the list of things to consider in the future.

      Delete
    6. I just wanted to add my two cents... since I'm one of the "lucky unskilled" player many seem to be griping about, currently sitting in 2nd place with fewer than 100 matches played.
      I have been playing cribbage on this app for many years and have reset my account a couple times as my game has evolved... so comparing my CR results with Classic or Single player would be pointless.
      Also, when playing on Classic my style of play varies from casual to serious to experimental to I-don't-give-a-f***.
      But on CR, I take it more seriously, concentrating on each decision, and playing consistently in the style I've found that works for me.
      I too have experienced both tremendous streaks of good and bad luck over the years. I was amazed with my 14 match winning streak early on in CR followed shortly with another 10 match streak.
      Being at or near the top means two things, each win will result in very few points gained and a loss results in a relatively larger number of points lost. So where is the incentive for me to play more matches? Other than the minimum two per week.
      Question... what is the penalty for not playing the 2 match/week minimum?

      Delete
    7. @Flyangler - Thanks for providing your perspective. Your question about incentive in playing more than the minimum is certainly valid. I would be interested to hear any ideas you may have about ways to improve that experience. Please email them to us at support@FullerSystems.com For your question about the "pnealty", there is no set CR point value for that, as the CR you see is the output of a relatively complex formula. Instead, the system adjusts one of the parameters of that formula (the confidence factor) for every week that you don't play. This makes the system less confident about your rank, and results in a lowering of CR.

      Delete
  28. I've played about 100 games (superdow101) and twice a game was terminated because "play time exceeded". Yesterday was the second time and I don't believe I had gone beyond the 10 second limit even once. The message said that win, draw, or forfeit would be based on score at the time the game was cut off. Is this a bug or just my imagination?

    ReplyDelete
    Replies
    1. The server will determine who exceeded the time limit or who disconnected to determine the winner. For specifics of your account and a particular game, please email us at support@FullerSystems.com so we can investigate it further.

      Delete
  29. I would like a manual/muggins league. True cribbage is counting for yourself and capitalizing on your opponents mistakes.

    ReplyDelete
    Replies
    1. Hi Sam, certainly something we are considering for the future if we can get enough interest in it. We need enough players who want this option before we could have enough to make the matchmaking work.

      Delete
  30. I have had multiple players purposely quit the game after I won the 1st of 3 matches. Yet my profile does not reflect those games. However, I have been disconnected twice; those show as loses.

    ReplyDelete
    Replies
    1. Hi, if you have a specific game for us to review, please email us at support@FullerSystems.com so we can investigate it further with you. What you are describing with the disconnects not counting as a win for you sounds like you may be in the Placement Matches phase still, where the games have to be played to the very end to count towards your 5 placement matches. They will still be in your stats.

      Delete
  31. I just played a match against Webby where that player exceeded the time limit no less than 10 times. Yet we kept playing. .......ss.ss s.s.s o.o.o.o.oo. s....l....o....w
    This "competitive " games seem to offer what you can't deliver.

    ReplyDelete
    Replies
    1. Hi, if you have a specific game, please email us at support@FullerSystems.com so we can investigate it further with you. Usually when you see something like that, it is a connection problem of some kind (possibly on your opponent's side). The game servers always are the system of record that can see both players and monitor the play time limits. There are several factors it considers in that, including connection delay. Again, we can definitely check any specific examples you can provide, just email us the details like your username and the time it happened.

      Delete
  32. Is the number next to my opponent's name their rank or CR rating?

    ReplyDelete
  33. So I believe you when you say you strive for fairness. I've enjoyed playing this game app... as well as been aggravated seriously many times. A background on me, I do risk assentment of uncertain events for the government. That said, in true randomness the likelihood of repeating highly improbable events says it's probably not random. Here's my classic case from playing. Dealt a 7,9,J(spades) and 3 Ks. Threw the 7 9 in their crib. Thought a Q of spades would be optimal and it was cut! 3 hands later on my deal the same hand was dealt and i threw the same 7 9 into my crib. Unbelievably the same Q of spades was cut to match the same J in my hand. That causes a huge problem in my mind. I could go on about a straight flush in diamonds with thr cut and my opponent had diamonds flush too and 3 cards also added to my straight... or 2 out of 3 hands caught inside cuts to match pairs of 6s and 8s while the 3rd hand had three 6s and a 7 with cut off an 8. Those are those incredible cuts others herr talk about. That game eended as a double skunk and left me shaking my head in disbelief. A few other strong trends keep coming up. So the algorithm must have some flaws to me because I believe your hard work to advance this game is evident. I do know from your write up on computer games that it has slanted hands on the brutal setting to test player's skill. So I know hands cash be skewed. Not sure what to believe on the inner workings and the apparent patterns i see and the striving for fairness. Thx...

    ReplyDelete
    Replies
    1. Hi, the way we determine if something is random involves more than looking at repeating of "highly improbable" events for a single player. Things must be looked at in aggregate over very large samples that respect the fact that we are talking about a 52 factorial sized set of probabilities. This is a huge topic, so I won't cover it here in a comment, but please read through the published audits we have done and explore why we choose the method we do to show the fairness and randomness of the shuffle and card distribution in games. With respect to the cut card, the system absolutely will always, without any changes ever, cut the card that is selected from the spread out deck shown. If you touch/select the 12th card in, you will always get that 12th card from the shuffled deck. The system doesn't alter that, or care about it in any way. There is no code written anywhere in the system that will attempt to alter the cards in any way. You also mention "the algorithm", but there is no mathematical algorithm in play when it comes to shuffling and dealing the cards out. It is just a random shuffle that, although you could consider a sorting function to be an algorithm of sorts, it is not programatically selecting the cards for you - it is just a random assignment and sorting that is done.

      I'm not sure what you are referring to with respect to a write up on the computer opponent games indicating that hands are slanted in any way. We absolutely do not do that. The single player game is just as fair and honest as the multiplayer games. It uses the same exact system of random assignment and sorting to shuffle the cards, and absolutely never alters the cards in any way. The difference in difficulty levels is just a matter of the amount of computation the computer does to determine discard and play cards. Never anything else. No exceptions.

      Delete
    2. But don't you love all of the 5's showing up on EVERY hand?
      J/K.
      I hope they fix this soon.

      Delete
  34. When does the weekly time period begin when a player must play two games a week? On a Monday? Thanks

    ReplyDelete
    Replies
    1. The clock ticks over just after midnight UTC on Sunday of each week. This shouldn't generally matter for the most part, as if you regularly play two games per week, it will not matter which day/time you choose to do so.

      Delete
  35. I was so excited today when I found this new way to play. It is so bad though.. 90+ % of the hands are double run for 8. I've never seen so many 5's!
    At least it is consistent to where in the rare case I score less than 10 points, the other player also scores less than 10 points. Otherwise it's usually 20 points each. It's not fun that way. I'll go back to the single classic game style to where it is more realistic until this is fixed.
    Please fix asap. I look forward to playing this new format

    ReplyDelete
    Replies
    1. Hi, we have not changed anything with the game itself when introducing this new feature. This simply adds a new way to rank players done through matchmaking. All the shuffling, dealing, and all other game play is identical and was not changed in any way whatsoever. Not even a single word or letter of the core game code was altered to add this feature.

      Delete
    2. I know, you would never do something on purposen to mess with the Integrity of the game. But I assure you it is off. Figure it out and fix it please.

      Delete
    3. First, understand I'm not trying to dismiss your experience as if it didn't happen. Strange things happen all the time in the world of random card shuffling, and I would have to review every one of your games in detail to know it for myself, but I'm fine assuming what you said did happen. Still, we do regular audits to make sure the shuffle and deal are always working correctly and have not been compromised in any way (we publish those sometimes too to the blog, and have never found any problems but we do them anyway), and we will absolutely do more, but we also use tools that track all changes made to every piece of the system (every letter/line of code for all systems). The areas that handle the shuffling, dealing, etc. are all shared with all types of the game (there is no way to change it for this competitive matchmaking system and not also change it for all other versions of the game - it is completely shared) and those code areas have not changed for many years. If they were to change, the code tracking system would immediately show where the code was changed and it would include a comment about why the change was made. These systems are called "version control systems", and we use them for everything. Certainly bugs can happen (but again, it would be a bug in both/all types of games), and this is why we do testing and after release audits, but it is essentially impossible that what is now called the "classic" games and the new competitive match games have different card shuffling systems. The only chance that could happen that I could even imagine would be something truly improbable like a hacker breaking into our code, changing it without our knowledge, compiling and deploying it to our servers without our knowledge and somehow now hiding it in such a way as we can't detect the change (for which every step along that path we have measures in place to both block and alert us of such activity). As a person of science I can't say that is 100% impossible, but it sure is extremely unlikely. It would be similar probability of solar flare activity causing it. Sure, it is possible, but so very highly unlikely.

      Delete
  36. I think it would be nice to have a one time rematch option. For those times when you feel as if you just got pummeled by the hammer of random. I believe it would add a bit of satisfaction to some games. And I'll put my vote in for 15 seconds during discard. Thanks for the app. I love it.

    ReplyDelete
  37. Hi again.....Well,,,,,,,,I will weigh in now that I have one match under my belt. Here is my take thereon. My goodness, 0nly 10 seconds per move??? That I find is sooooo fast. One cannot even begin to think what one's best play should be. For the 3 games of my match, I had 6 warnings! :):) So yes, if the games are for very fast players, who do not need thinking time, then it is fine. Heck, just running my finger over the computer pad took some time. Now, this first match of mine was against a fine player I'm sure. He ripped me on the first game (skunked) but I got the next two - including A SKUNK back at him. :):) But,,,,,,,, to do so, I simply let go of any hope of using a strategy and blasted away with "not thought about" cards that "seemed" to be the correct one's to play. If I goofed, and had to change my mind, up came the dread "hurry - up" message :):) My "take" is that I just did not find the match to be fun - given the speed. But, perhaps that is your intention? If so, so be it. I have plenty of fun using your other options and do appreciate the amount of work you have expended in creating what we have so far. See ya. Jerry

    ReplyDelete
    Replies
    1. Hi, yes, we do understand that 10 seconds may be too fast for some and ultimately that means this style of play is not for you if you can't have fun with it. We may add some options to the competitive style of play in the future to open it up a bit more for others.

      Delete
  38. Why is it considered a competitive match when opponents are often over 1000 points lower than me? I would not choose to play someone that much lower if allowed to choose

    ReplyDelete
    Replies
    1. Hi, we would have to look at your specific examples to know the exact reason for the matchmaking algorithm doing what it does. The important thing to remember is that your CR value is a calculated value based on several factors, and the matchmaking system is considering all of those factors (and other things like who you played recently) when making a match. Not all matches will be exactly your level, and some will be above or below. The system works to test out your rank in multiple ways.

      Delete
  39. Ya know, cards are amazing. The combinations are endless. BUT I'm having a problem. I went 16 games in a row in wins. Seemed a little to good to be true, maybe lucky. I made the top 20s, great, awesome, feels good. Since then I have gone 10 wins and 50 losses. The disparity has been very frustrating and frankly, impossible. I've had way too many 0 hands. Somethings up.

    ReplyDelete
    Replies
    1. Hi, it sounds like you have definitely had a mixed bag of luck as a contributing factor. We have worked very hard to keep the game fair and honest in every respect and regularly audit the game for the same. I won't go into all of that here, as it is covered extensively already in the game Help/FAQ and the multiple blog posts on the topic.

      Delete
  40. How many total players have a CR? The top 50 is more impressive if there are 5000 players vs 500. That should be an easy number to display on the leaderboard.

    ReplyDelete
    Replies
    1. Hi, we don't share specific user numbers, but there are many several thousands of players who participate in the Competitive Matchmaking system. Our next update will include a percentile ranking to help put a perspective on this for you.

      Delete
  41. Why is it considered a competitive match when often the opponent you match me with are over 1000 points lower than me? I would never choose to play someone that much lower!!

    ReplyDelete
    Replies
    1. This looks like the same question you already asked, so the same answer applies here.

      Delete

Keep all comments clean and appropriate, and only post if it moves the conversation forward. All others will not be published. If you have a support request or general comment, please email us at support@FullerSystems.com