The Ratings Committee (RC) was involved in a number of projects this past year, though unfortunately some of our work was hampered by the inability to test proposed rating system modifications due to resource constraints at the USCF office. Our focus this year was on the final details of the new title system whose implementation is imminent, clarification of match rules, and rating system monitoring tasks. A few other minor issues were also addressed. Most of the work on the title system has been completed, but a few small issues remain. The basic premise of the title system is to award permanent titles based on sustained performances at particular rating levels. As an example, a player who is vying for the 1800-level ("1st category") title would need to demonstrate several qualifying tournament performances in which his/her game results would be considered sufficiently impressive for someone rated 1800. For each qualifying performance, a single norm is awarded. Once five norms are collected, a title for that level is issued. Norms and titles cannot be lost through poor performance or inactivity. A player may be working on several titles simultaneously. In the latter half of 2009, we worked with Mike Nolan to retroactively apply the title system going back to 1991 to examine the distribution of titles earned. We found that roughly 20-30% of players at a rating of near X earn the X title, and about 50-60% of players rated X+100 earn the X title. As an example, this means that players around a rating of 1600 would tend to have a 20-30% chance at eventually earning the 1600 ("2nd category") title, whereas players rated around 1700 would have closer to a 50-60% chance of earning that title. Unlike the title system from the early 1990s, it is intentionally more challenging to earn titles in the current system. One detail of the title system that is still under consideration is how to handle situations where a player does not play four or more games in events of four or more rounds. The current rule states that players may only earn norms from events of at least four rounds, but the ambiguity is whether all such games are played. In general, there is agreement among RC members that if a player has played four games, then only those games would count toward the norm computation. In cases where fewer than four games are played, one possibility is to restrict norm eligibility to situations in which a player actually plays four or more games. In this case, players who do not play at least four games are not eligible for norms. An alternative idea that has been proposed is to consider a player eligible for a norm if he/she has played fewer than four games in an event of four or more rounds as long as one of the unplayed games is a full-point bye or forfeit win, indicating that the player had intended to play a game. In such an instance, a norm would be awarded (at the relevant rating level) if the played games were sufficiently impressive. In testing this latter approach on events from 2009-2010, only one player benefited from this approach, suggesting that allowing norm eligibility in four-or-more round events when fewer than four games are played does not measurably change the overall frequency of awarding norms and titles. This issue will be resolved shortly. Over the past year, the USCF office had to confront many instances of tournament submissions that contained multiple games among player pairs (which generally should be rated as matches), and unusual match submissions. One such match submission involved a provisional player competing against an established player. After some RC discussion about whether it is appropriate to rate matches between established versus provisional players, we discovered that unambiguous match rules were in effect since 2006 that precluded rating matches between provisional and established players. The rating system specs had not been updated to reflect this rule which was the source of the confusion. The document has since been updated. In a related matter, the specs have been modified to describe the current method of updating ratings based on FIDE events that are not directly USCF-rated. The specs can be downloaded from http://www.glicko.net/ratings/rating.system.pdf The Executive Board (EB) presented several items for RC consideration stemming from their November 2009 meeting. They asked the RC chair about whether rating floors could be implemented at ratings of 1200 and 1300 (currently the lowest floor above the absolute floor of 100 is 1400). While the RC chair reminded the EB liaison that rating floors can only distort the validity of ratings, he recommended that if they were implemented then the RC would monitor whether these two new floors cause any noticeable shifts in ratings. The EB also proposed a new "victory points" system, which is an award system much in the spirit of "master points" in bridge. While the RC generally had no problems with the details of such a system, the sentiment was expressed that having both a victory points system concurrently with a title system might dilute general enthusiasm for both systems. The EB further asked the RC to consider methods for connecting the Quick Chess (QC) and regular rating systems, and to study whether the K-factor in the established rating formula needs to be larger for high-rated players. These are both issues the RC has been considering for awhile, but to fully evaluate how to proceed requires the use of computer resources at the USCF office that have not been available for ratings testing. Both of these issues are high-priority for the RC, and once the USCF backup server is available to run simulations these tasks will be carried out. One minor issue that was brought to the attention of the RC chair was an instance where the "special" rating formulas for handling provisional ratings produced unintuitive results. The situation involved an unrated player who had perfect scores in several small events. The problem is that when a player has a perfect result in a first tournament and gets a provisional rating based on N games (often the largest of the opponents' rating plus 400), the second event treats the post-tournament rating from the first event as though it were N draws against players of that rating. This could conceivably result in strange results, especially if the opponents in the second event are much stronger than those in the first event. The reason for treating the results from the first event as N draws is to avoid saving the actual game results. While the particular situation that was brought up is certainly rare, we will investigate whether it is possible to save the actual game results in the rating program to avoid these odd behaviors. Each year the RC performs a set of diagnostic analyses to monitor trends in the rating pool. Overall rating levels have deflated from the mid-1990s through 2000 when rating floors were decreased by 100 points without a counteracting inflationary mechanism. With the new rating system implemented in 2001, ratings started to re-inflate. The RC has the goal of restoring rating levels back roughly to where they were at the end of 1997. The focus of RC work has been on players with established ratings who have been active over the current and previous three years and who are aged 35-45 years old in the current year. Based on the continued decline in ratings for this group, the RC recommended two years ago lowering the bonus point threshold from B=10 to B=6 to accelerate the re-inflation of the rating pool. This change was applied retroactively to the beginning of 2008. Over the past two years, the average rating for this group has increased 20 points in 2008, and about 25 points in 2009. The average rating is currently 40 points lower than the average rating at the end of 1997. A concern was expressed among some RC members that leaving the bonus point threshold at B=6 for another year might result in overshooting our target, since the effect of lowering the threshold not only helps to increase players' ratings directly through bonus points, but indirectly to the opponents where the benefit of the bonus point increase takes longer to observe. We will be working in the next few months with the USCF office to test various bonus point thresholds retroactively to examine the "trickle-down" effect of increasing B to a larger value. Depending on the results of this analysis, we may recommend increasing B to either 8 or 10 in the next few months.