|
|
Tournament ratings were introduced in Fall 2004. Ever since then, they've been the basis for much discussion, debate, celebration, commiseration and even occasional controversy. Love 'em or hate 'em, they seem to have made things more interesting. See the Background below for an explanation of the different ratings. Important note: Players who do not wish to be rated should refer to the Dr. Names section below. Outages: Ratings are calculated immediately after each tournament. Occasionally, however, the process fails, and it is not self-correcting. If you happen to notice the ratings are not current for a particular theme, please post a quick note to the Netcell discussion board and let us know. The ratings crew and your fellow competitors will very much appreciate it.
BackgroundWhile freecell.net has always offered historical data for the top finishers in past tournaments, until Fall, 2004 there wasn't a consistent way to put it all together and really compare skills with other players. The "Ratings Crew" designed a system to do just that. The rating system employed here is based on the same model used in many other competitive events, such as chess, although this model has been customized to our situation. It has been extensively tested, and is more flexible and robust than any of the Freecell tournament rating systems previously explored. Overall, NetCELL tournaments are well suited to a rating system such as this, and it serves Freecell players over a wide range of skill levels. BasicsRatings are based on not only your final position in each rated tournament you play, but also on who you placed ahead of and who placed ahead of you. That is, finishing ahead of a higher-rated player will improve your rating more than outpacing a lower-rated opponent. The rating attempts to predict where you will finish in a given tournament, and with each successive tournament played the ratings are updated based on your actual performance. Final elapsed time and number of wins are important only insofar as they determine your final standing relative to the other players; no special consideration is given for margin of victory. How do I get a rating?There are two hurdles to becoming a rated tournament player. First, for your appearance to count at all, you must, just once, complete a certain number of games in each theme you wish be rated, as follows.
Note: For the first almost-five years of ratings, players would not have new ratings calculated after any tournaments in which they failed to reach the above cutoff limits. As of June 15, 2009, however, completing just one game in a tourney triggered a ratings adjustment. Which tournaments are rated?Ratings have been calculated back to the very beginning of NetCELL tournaments. Separate ratings are calculated for each tournament theme. This is done to highlight players who excel in a particular type of competition. Masters tournament ratings are perhaps the most meaningful because the Masters draws the most competition; cumulative ratings are also provided but may not predict well how a player will do in a particular tournament theme. Ratings are constantly readjusted based on your most recent performance, and therefore no effort is made to exclude outdated tournament results. Because the rating focuses on a player's current form, one bad performance will not have a lasting impact. If, for instance, a 1700-level player has a poor outing one week, after two or three more characteristic finishes the rating should be restored. What does my rating mean?As a statistical model, your rating not only shows your relative rank in skill, but also gives some idea of how far separated you are from other players. Generally speaking, a difference in ratings of 200 points represents one standard deviation, meaning the higher-rated player should place above the lower-rated player approximately 68% of the time. If the difference is 400 points, the higher-rated player should win 95% of head-to-head matches. A 600-point difference in ratings indicates a 99% dominance. Can you explain "rating" and "ranking?"Your rating is the figure that is calculated based on your combined performance in all rated tournaments. However, the rating figure by itself is significant only in relation to other players. So the calculated rating is used to determine where you rank among your peers. What is a "good" rating?Many players who participate in tournaments do not achieve a rating at all, so in that sense all ratings are good. Rated players begin with a rating of 1500, and this number can fluctuate up and down widely, especially in the early going. (Note that no direct comparisons can be made to a rating figure from other similar rating systems, such as those used in chess or in Yahoo! games.) As ratings rise, this volatility decreases proportionally. This allows newer players to quickly move to their proper place in the list, and lessens the damage on a player's rating after one uncharacteristic performance. Player names appear in color in the ranking lists based on tiers of ratings, as follows.
What do the green and red arrows signify?Over the course of time, most players will complete some tourneys where they perform either far better or far worse than their rating would predict. In such a case, we delay giving or taking away the full points earned or lost. We do this by calculating two ratings for each player, one raw and one smoothed. The smoothed one is the one that appears on our ratings pages, but you can view your unsmoothed rating by hovering your mouse over the smoothed one. The difference between your smoothed and unsmoothed ratings is your momentum. If it is 24 points or more, a green arrow is displayed, but if it is -24 points or less, a red arrow is displayed. In general, new players moving up towards the "true" rating will be the ones most likely to have a green arrow. Keep in mind that the smoothing has no long-term effect on your rating. You end up in the same place - the smoothing simply takes out some of the bumps along the way. How can I improve my rating?Click faster! Part of the fun in seeing your rank will be to see which players are closest to you in skill and focus on outplaying them. One important consideration, though, is that casual play in a rated tournament will be treated as a serious effort, and therefore should either be avoided, or an alternate name used. Over time you will most likely find your rating settles into one place unless your skill level changes dramatically, so focus on incremental improvement. Dr. NamesIn 2006, the Ratings Crew introduced "Dr." (don't rate) playing names in the interest of increasing the accuracy of the tournament ratings. The concept is simple: any name that begins with "dr." (or "Dr." or "DR." or even "dR."; e.g., Dr.Bombay, dr.calicokid) is ignored for ratings purposes. If you start late, expect to be interrupted, or are temporarily technically or mentally compromised in some way, there are two reasons to play as a Dr.: 1. To protect your own rating; and Of course, you can always play with a Dr. name any time you don't wish to be rated, for whatever reason. We also ask people to choose one and only one non-doctor name per theme. For the most part, this has worked very well, and we have been pleased with the results. Thanks very much to all who have cooperated with the doctor system. We hope you enjoy ratings! The Ratings Crew,
|
| |||||||||||||||||||||||||||||||||||||
