23 September 2017 Auckland Scrimmage Summary
In our seventh In The Zone event for the year, we had 31 teams show up to put their robots to the test. This is the highest turnout we’ve had this year, although the last couple of times we’ve had 28 robots so it’s not much of an increase. Each team played 6 Qualification Matches, before heading into the Alliance Selection.
At the end of the Qualification Matches, 7682E (Wingus & Dingus) secured the top spot, with 2941B (Oats Robotics/Otumoetai College) in second, followed by 2915C (Lynfield College) in third. All three of these teams had won all their Qualification Matches. In terms of Autonomous Points, these three teams won 5/6, 4/6 and 3/6 of their possible Autonomous Bonus opportunities respectively. The interesting thing to notice with these rankings is the difference from the last scrimmage, where SymbiOHSis/Onehunga High School (Team 2900) had taken out all three of the highest ranked teams.
The finals were played between the first and seventh seeded alliances. These composed of teams 7682E & 2900D vs 2900C & 2900B. The first seeded alliance won the first game with a score of 93-76, followed by a second score of 77-56 to win the event. Congratulations Wingus & Dingus and SymbiOHSis/Onehunga High School!
Now for what the numbers tell us about the event overall:
Perhaps the feature most eye-catching right away is the significant difference between qualification and elimination matches. For the single alliance match scores, we’re seeing a median of 47 with a maximum of 111 points for the qualification matches, or a median of 77 with a maximum of 101 points for the elimination matches. There’s a massive difference in medians between qualifications and eliminations here (47 vs 77), although intriguingly the maximum is higher for the qualification matches. There’s a couple of reasons why this could exist – firstly, the randomness of the qualification match alliances could have had two particularly strong robots against two weak robots, or perhaps in the elimination games there was just more defense. We’ll consider this second idea again later on. For the combined alliance match scores, there’s a median of 98 and maximum of 153 points for the qualification matches, and a median of 144 with a maximum of 184 points for the elimination matches. Again, an enormous difference between qualification and elimination games, with the medians of 98 vs 144.
Compared to the last scrimmage, all of these medians are slightly higher than last time (some more than others). Not all of the maximum scores have improved, but it’s a little hard to expect an improvement every time that when those are the more “extreme” games.
We’re seeing the points scored in games coming in similar proportions to the previous event. Between qualification and elimination matches, the proportions are roughly the same, although the number of points scored in elimination matches is much higher as discussed previously. Most of the points are still coming from the mobile goals (~60-65%), with cones coming in second (~20-25%). Highest stack bonuses only contribute about 11-12% of the final score.
The most popular place for mobile goals to finish up was the 10pt zone, in both qualification and elimination matches. We’re seeing just under half of all mobile goals in the 10pt zone for the qualification matches, and a bit over half of all mobile goals in the 10pt zone for the elimination matches. Whilst we might expect from the above figures to have about 3 mobile goals unscored at the end of a qualification match, we only expect 1 mobile goal to remain unscored at the end of an elimination match. Hopefully we see that 36% of unscored mobile goals in qualification matches reduce soon. It’s worth keeping in mind that for the 20pt zone, the maximum possible percentage of goals in a 20pt zone is 25% (you’re only able to have one mobile goal scored in each 20pt zone).
We still have yet to see many cones being stacked in games. For the qualification matches, there’s a median of 9 cones stacked per game, and for the elimination matches a median of 16. Hopefully we see this pick up soon – 16 cones between all robots on the field is only 20% of all cones scored. Most stacks are still only up to a few cones tall, with the highest stack being 7 cones.
Here we see that those highest stacks of 7 cones are coming from mobile goals in the 10pt zone (for both elimination and qualification matches), or from unscored mobile goals (only for the qualification matches). We’re seeing stacks of up to 5 cones on the stationary goal, although it’s probably fairly safe to say that most of the focus is going into the 10pt mobile goals, as discussed throughout this summary. The highest stacks in the 5pt zone last time were 7 cones high for qualification matches or 5 for the elimination matches, but now just 3 cones tall for both qualification and elimination matches. That’s probably not much of a concern when so few mobile goals are being scored in the 5pt zone right now, with such low numbers of cones being stacked.
Other details:
- Highest single alliance score of the day: 111 points
- Highest combined alliance score of the day: 184 points
- Highest stack of the day: 7
- Qualification matches with an autonomous bonus awarded: 24 (51.1%)
- Qualification matches with a match affecting autonomous bonus: 9 (19.1%)
- Median difference in qualification match scores: 33 points
- Highest driver skills score: 72 points (achieved by 7682E and also 2900C)
- Highest programming skills score: 7 points (achieved by 7682E)
- Highest combined skills score: 79 points (achieved by 7682E)
Finally, looking into how we’re progressing through the season brings up a few interesting points. We’re seeing a steady increase in scores across the board for the qualification matches, yet there’s a decrease in the median winning score (and maximum score) for the elimination matches. One could suggest that as the weaker teams are improving, they’re making it gradually more difficult for the stronger teams to maintain a significant lead and perhaps are as a result reducing their score, however considering that so few cones are actually being used, the argument of a battle for access to cones doesn’t really hold up, and we know that the mobile goals are still being scored. Instead, it’s much more likely based on the numbers we’re seeing that defense is becoming more popular, and perhaps the stronger robot’s time is being wasted a little more. Despite the gap between winning and losing scores increasing slightly for the qualification matches, it’s reassuring to see the growth and improvement that’s actually occurring.
Thanks to all my fellow volunteers for lending a hand with the scrimmage. Thanks also to Kiwibots and the University of Auckland for organising the event.
The next scrimmage is in just under three weeks’ time (14th October) at Kristin School. This is the last scheduled scrimmage before the Asia Pacific Championship (APAC), so it may (hopefully!) be a little more competitive. However, there is also a 7-week gap between the next scrimmage and APAC, so I’m sure there will be some rebuilds and new robots popping up over that 7-week period.
