Announcement

Collapse
No announcement yet.

Ranking Points - should something change?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Ranking Points - should something change?

    At the recent Alabama event the weaknesses of the ranking points system became very obvious. We were about to start our 4th match out of 5 matches. We had won all our matches, but had very little ranking points at that point. In one of our first 3 matches the opposing alliance robot tipped over shortly after teleop started and the other robot did not score very many points. In another round both robots did not score basically any points as one got disabled and the other hit its power switch on the wall right after autonomous because it jammed itself between the wall and the balancing stone and had a bad location for its power switch. The final match we were going to be on an alliance with one of the top 2 robot teams (if not the top team) and facing opponents that had not done well up to that point.
    Our average alliance points would easily place us in the top 3 teams as long as we won our matches. There were more than 5 undefeated teams at this point. So the only way we had of winning a seat as one of the top four was to increase our ranking points. The only way to advance to regionals was to be the inspire award winner or be the winning alliance captain, which required being in the top 4 during qualifiers. So the natural question is "Do we put glyphs into the opposing alliances cryptobox to increase our ranking points?" With our autonomous and the alliance partners points we could have won even if we build the opposing alliance a column and we would have enough ranking points to make it into the top four. So we debated back and forth about whether we should do it. In the end we decided it may be humiliating for the opposing alliance to have that happen to them, so we didn't do it and we ended up in 5th place over all, 17 points behind the 4th place team. The 4th place team was way behind the 3rd place team in ranking points, even tho they had one of the best robots.
    Should we really have a ranking point system that makes us contemplate the options of A) losing even tho we won all our matches, or B) humiliating the opponent in order to win?
    The other option is to write a program that purposefully knocks the wrong jewel off so the opponent gets points. As long as the opponent does not know you are doing this, it may not be humiliating to them.
    Should we really have a ranking point system where we have to write programs to purposefully hit the wrong jewel so that we can win the robot game?

    I would love to hear how other teams would handle this.

    We are rather new to FTC, so maybe we don't understand how the ranking points system works. I am not sure why your average alliance score would not be factored into the ranking points, especially for a competition like this years where you really don't interact with the other alliance for anything. Last year when anyone could change the beacon color, or could pick up the limited number of balls, how good your opponent was could impact how well you could do, but that aspect was completely eliminated this year. Especially since just about any interaction with the other alliance gives a penalty.

  • #2
    RP isn't a great tiebreaker, but the real problem is that your regional championship should be running a lot more than 5 matches per team to prevent stuff like that from happening. I've seen the stats on this before someplace (I will post them if I can find them), but you need at least 8 or 9 matches per team at an event with 32 teams before the seeding process becomes meaningful.
    FTC 4962 / 3638
    FLL 11 / 21 / 9293

    Comment


    • #3
      Also, I'd advocate for the type of ranking system now used in FRC, where you can get bonus points in the standings for specific achievements, and then tiebreakers in the standings are decided by other match achievements. This year for Power Up the tiebreaker is (1) cumulative parking/climbing score (2) auto points (3) ... etc.). It's a nice combination or recognizing the achievements of each alliance while still giving individual robots some reward for being good even though their alliance partners may be limited.
      FTC 4962 / 3638
      FLL 11 / 21 / 9293

      Comment


      • #4
        We have had a lot of conversations about RP points and figuring our strategies going into tournaments. I like the RP system because it does base your scores off of your strength of schedule. You could be a 1 loss team and have a lot of RP points because you lost to a team that threw up more points than your high score. I think this has to play into the strategy which is why it is legal to score in an opposing alliances cryptobox.

        We have considered scoring for the other team and would do it if we had to for the RP points. I don't view it as humiliating to the other team as it is just a strategy. We have also discussed situations where you may face an alliance that you know is going to win the match. In this case, not scoring any points, to not give them any RP points could drop them out of 1st place. In this situation we wouldn't get RP points either but you may have a partner that could end up #1 that would pick you knowing what you did helped propel them to 1st.

        This is what I love about FTC and all of the strategy that goes into it and all of the scouting and data collection our team gathers at tournaments. There is definitely a lot more to it than just robots!

        Comment


        • #5
          Originally posted by BSV View Post
          but you need at least 8 or 9 matches per team at an event with 32 teams before the seeding process becomes meaningful.
          That is good to know. If we would have had that many rounds, there is a good chance some of the top teams would have gotten a loss, and their QP's would have gone down. It definitely would have been a good test for robustness of the robots.

          Comment


          • #6
            I have also thought that the RP system should be changed. You have absolutely no control over the points the other alliance accumulates. That is, without crossing over and helping them out. I believe the autonomous portion score should weigh in heavily on the RP. Perhaps some type of weighted average could be used. ((Opposing alliance score X 50) + (Own alliance autonomous score X 25) + (Own alliance total score X 25)) / 100. Something to that effect would keep part of the current theory of more points for beating higher scoring teams but also give some strength to actions the teams can control.

            Also, I have thought that 5 rounds is too few. A ball rolling into the cryptobox on autonomous, causing the glyph to fall out. A mechanical malfunction of some sort. Lost connectivity. Your alliance partner's robot being disabled due to not fitting the 18 inch requirement. There are many factors that can go wrong, many of which are random events. It looks to me like more rounds would reduce the randomness and be a truer indicator of the quality of the robots.

            Comment


            • #7
              In the past, our team has intentionally scored for the other alliance to build up our ranking points when we knew we had a big lead. We have stopped doing that now since it doesn't seem right to risk possibly losing, and really isn't a great game play strategy. On another note, I do like the game play with autonomous sensor challenges, it is a great display of the technology for spectators, and the programming software is also so much better, and the field elements aren't as huge as FRC, much more manageable to practice with now, hope to see all of that continue.

              Comment


              • #8
                Hi Folks,

                On the topic of Ranking Points:

                While the ranking point system isn't perfect, it does reward teams for winning matches that are more difficult. A match won against a very competitive alliance should be more valuable than a match against an alliance that has a bad match. Because this is an alliance based competition, how well you work with your alliance partner is important. This is why alliances receive points together, and why teams are not rewarded individually for their contribution. While not perfect, this is a simple way to reward difficult tasks, and promote teamwork between alliances.

                JoAnn

                Comment


                • #9
                  Hi Folks,

                  On the topic of the number of matches per team at each event:

                  FIRST has set a hard bracket of matches at Qualifying Tournaments, Meets, League Tournaments, and Championship Tournaments with a low number of 5 and a high number of 6 matches per team. The reason for this has much to do with fairness and logistics.

                  Setting a consistent number of matches across the world means that teams in Virginia can expect a similar experience to teams in British Columbia. This seems reasonable to us.

                  Tournaments are very long days, and forcing more matches into an event means that the cycle times need to be faster, and the day needs to be longer.

                  When cycle times are faster, teams have less time to repair their robot, less time for strategy, less time to scout, and less time to take advantages of the social aspects of being at an event. This makes the entire day stressful for everyone.

                  Faster cycle times also impact the event volunteers. Referees must work faster, scorekeepers must work faster, and the FTAs, have less time to devote to individual teams who really need their help.

                  Ultimately, we are concerned about the quality of the events, and the quality of the event experience for all of our participants, from the rookie who might be struggling, to the uber team who rises to the highest level of competitive ability.

                  JoAnn

                  Comment


                  • #10
                    In VEX they sort of fixed this by introducing another ranking mechanism call AP's (Autonomous Points). So teams are ranked first by WP's then AP's and finally SP's (SP's being the VEX equivalent of RP's).

                    Teams that excel in their autonomous routines are rewarded through better ranking and rightly so.

                    It would be great to see HQ put together a survey to gauge the level of interest in changes to qualification match counts, possible RP calculation changes, and other ideas such as introducing AP's. Hearing "that's how it is" from HQ is fine however teams themselves need to have a say too and there's plenty of untapped knowledge out there.

                    Comment


                    • #11
                      This is my first year being involved with FIRST. The following quote from response #9 sums up perfectly my impression of every event my son's team participated in. In each case, getting the day completed and going home seemed to be the first and foremost priority for some of the officials. Time for the things mentioned was minimal. The rush to push everyone through the process as fast as possible did indeed make the entire day stressful for everyone. The push for speed also, in my opinion, was a major contributing factor in errors made by scorekeepers (TWO data entry errors on one round), field resets (Pictograph placed centered on the middle of the balancing board instead of offset by 3 inches), etc.


                      "When cycle times are faster, teams have less time to repair their robot, less time for strategy, less time to scout, and less time to take advantages of the social aspects of being at an event. This makes the entire day stressful for everyone.

                      Faster cycle times also impact the event volunteers. Referees must work faster, scorekeepers must work faster, and the FTAs, have less time to devote to individual teams who really need their help."

                      As a father, I see the many hours that my son and his teammates put into the robot. They spent dozens and dozens of hours researching, building, programming, tweaking, practicing, etc. I also believe that the quality of the event experience should be a positive one. It seems disrespectful to the amount of effort put into the process to just rush everyone through as fast as possible.

                      So, I agree with you that the fast pace causes undue stress. Perhaps there is a way to build in some time somewhere for those pieces. Maybe a couple of minutes between rounds. Maybe a 5 or 10 minute break each hour specifically set aside for networking, etc. There are many options and I believe it would greatly enhance the quality of the events.

                      Also, the small number of rounds seems to leave a lot to chance. More rounds would give a better indicator of the quality of the robots since it would diminish the luck factor. I respect the challenge of finding the right balance. Again, as a parent, I have spent a good deal of time driving to and from events. Adding a little time to the event would be very worth it from my point of view. Moving from a minimum of 5 rounds to a minimum of 6 rounds is a 20% increase. For a 4 hour event, this only adds 48 minutes. For several events, I drove more than 48 minutes each way to attend the event. That is not a significant increase in the overall time commitment. However, it would reduce the luck factor by 20%. This would give a truer indication of the quality of the robots.
                      Last edited by Mentor 777; Yesterday, 05:48 PM.

                      Comment


                      • #12
                        Originally posted by JoAnn View Post
                        Hi Folks,

                        On the topic of Ranking Points:

                        While the ranking point system isn't perfect, it does reward teams for winning matches that are more difficult. A match won against a very competitive alliance should be more valuable than a match against an alliance that has a bad match. Because this is an alliance based competition, how well you work with your alliance partner is important. This is why alliances receive points together, and why teams are not rewarded individually for their contribution. While not perfect, this is a simple way to reward difficult tasks, and promote teamwork between alliances.

                        JoAnn
                        Just wanted to point out that this system does not reward teams that because of luck of the draw got to play with weak teams against stronger teams. We had most RP in our competition. In fact almost double of the RP of the 1st alliance Captain team. That means way to much left to a chance.
                        I'd suggest implementing OPR would be a great improvement to the leader board standings. Looking at our event results on http://ftcstats.org/, it reflected team strength much more accurately that QP/RP

                        Comment

                        Working...
                        X