[FiM] Week 3 Recap and Week 4 Analysis
Week 3 is over and this year’s competition season is finally taking form. We saw some great competition last week, an increase in partner hang consistency, additional autonomous modes demonstrating how valuable it is, and ultimately a better grasp of this years game on all levels. Now that most teams have played at least once, this week we want to dive into some updated data and draw what conclusions we have been able to see As always, our weekly post will focus on the data available and rely on subjective opinions as little as possible, however there is something to be said for the projected potential of a team as long as there is rational behind it.
All information we look at is based on the current most recently played week outside of our Weekly Rankings which will look at a team’s most recent event. We think this is important to note, since we all know that robots usually go through performance improvements between events, but a lot of documented stats can take the entire season into account. How is a team doing NOW, and how does that project into their future? How is the competition evolving NOW vs back in Week 1 when nobody really knew what they were doing?
This week let’s start with a series of interesting data points that we touched on in earlier weeks, and we can talk about what it means. Since the beginning of the competition season we have been talking about importance of scale ownership, how it is a must-have if you’re looking to be successful.
Based on Week 3 Events:
-An alliance that had more scale ownership time won 90% of the time
-An alliance that had more climb points won 80% of the time
-An alliance that had more switch ownership time won 78% of the time
-In the elimination rounds, an alliance with more scale points at the end of auton (excluding matches where neither alliance scored scale points in auton), had a 73% chance of winning the match.
-Average winning point margin was 157 points.
While this is good and interesting pieces of information…it really drives us to the same conclusions we’ve been talking about previously: Controlling the scale is the highest priority action you can do during the match. If you are better at controlling the scale than the opposing alliance and get control at the beginning of the match, you dictate the flow of the game and can control the game elements, and you’re going to win the majority of the time (duh). While your opponents are trying to take the scale back, you can spend time taking their switch, buffing your own switch, scoring cubes through the exchange, scoring even more cubes on the scale, or taking extra time to ensure the double climb. If your opponents lose control of the scale at the beginning of the match, you almost always have to take the scale and their switch to catch back up and win.
Last week we reported the Rest of World data to look at top scale ownership and top climb Rank Points (RP) robots. To be honest, while it showed that we have a few michigan teams within that top tier, we were not happy with the outcome and want to stay focused on Michigan. We also had a community suggestion come to our inbox that asked to present this information in an interesting or different way, so below is an graph showing teleop scale ownership percentage vs climb RP achieved percentage, then sorted by average overall RP for the teams from their latest event. We think that this gives an interesting, yet telling, understanding of the robots that should make up the top seeds at competitions down the road. Specifically, while there is no real strong correlation between avg. RP and Teleop Scale Ownership %, there certainly is a correlation between avg. RP and Climb RP %. This is important because like we have mentioned previously, with many scale bots in the crowd, it is going to be the climb RP that will push those teams to the top of the seeding rankings. Similar to 2017, that extra rank point will throw you to the top of the rankings, even with multiple qualification losses. We also feel that climb RP % will continue to increase as these teams get more efficient at achieving the climb, however, scaling ability is more likely plateaued, which should only make this advantage larger as we get to events such as MSC.
Supporting Data:
Team Auto RP % Climb RP % Tele Scale Own % Week Avg. RP
1 67 91.67% 83.33% 59.42% 3 3.25
2 2767 66.67% 66.67% 77.98% 2 3.17
3 5561 91.67% 50.00% 56.25% 3 3.08
4 3707 91.67% 41.67% 56.69% 3 3.00
5 3357 83.33% 16.67% 73.83% 2 3.00
6 33 75.00% 33.33% 63.42% 3 2.92
7 1918 50.00% 25.00% 75.71% 1 2.75
8 2474 83.33% 8.33% 53.53% 2 2.75
9 27 100.00% 50.00% 49.42% 3 2.67
10 5234 75.00% 8.33% 55.75% 3 2.67
11 5460 91.67% 33.33% 60.65% 3 2.58
12 245 33.33% 25.00% 70.70% 3 2.58
13 4362 91.67% 16.67% 71.30% 3 2.58
14 5674 58.33% 16.67% 67.10% 2 2.58
15 3538 50.00% 8.33% 71.81% 3 2.58
16 3604 83.33% 8.33% 66.59% 3 2.58
17 3618 75.00% 0.00% 58.24% 3 2.58
18 3656 50.00% 33.33% 67.17% 3 2.50
19 4391 58.33% 8.33% 71.73% 3 2.50
20 4003 66.67% 0.00% 89.12% 2 2.50
21 302 50.00% 0.00% 76.34% 1 2.50
22 217 50.00% 8.33% 69.75% 2 2.42
23 4405 83.33% 8.33% 60.29% 1 2.42
24 5152 66.67% 8.33% 47.90% 3 2.42
25 6090 58.33% 0.00% 76.73% 3 2.42
Some other fun and interesting, but maybe not so important nuggets from this data shows:
Only 5 teams across all 3 weeks have achieved an avg. RP above 3.00 (67, 2767, 5561, 3707, 3357)
- 33 had a 33.33% climb RP %, talk about taking your team number to heart!
- 27 is the only team so far across all weeks to achieve the auto RP 100% at an event
- Consistent detachable “hook and tether” style robots are having moderate success at getting the climb RP. While more reliance on your alliance partner is required, these teams should continue to do well as weeks progress and their partners are more able to consistently climb.
- 67 is the only team so far to achieve higher than 80% climb RP % across all weeks
This is the same dataset but filtered to show the top climb RP % robots in Michigan, surprisingly only 4 teams have a 50% or above climb RP rate. From this top 10 list we have a range of all of the different types of climb robots including Partner lift (ramp/forklift), hook and tether, and ramp bots)
Top Climb RP % Robots
Team Climb RP % Ramp Type
1 67 83.33% Buddy Ramp
2 2767 66.67% Buddy Ramp
3 5561 50.00% Hook and Tether
4 27 50.00% Buddy Forklift
5 3707 41.67% Hook and Tether
6 308 41.67% Ramps
7 33 33.33% Hook and Tether
8 5460 33.33% Hook and Tether
9 3656 33.33% Hook and Tether
10 3668 33.33% Hook and Tether
Weekly Rankings:
So since many teams have played once and quite a few teams have played twice, we finally feel we have a good grasp on which teams have the top robots in Michigan. After watching many many matches (so many omg) from across all events, our general ranking philosophy is based on two things (in order of priority):
- “Ability for the team to achieve the top seed at MSC”, meaning the teams who we feel will be able to get the highest avg. RP (i.e. great scale bots with consistent climb RP mechanisms)
- “Ability for the team to win the event by controlling the scale” meaning that into elims, a robot that has an amazing scale capability could out-perform a lesser scale bot that can climb.
-
2767
2767 and 67 are neck and neck for the top spot, however with 2767 yet to play their second event, we think any improvements may give them an edge. They have a quick and consistent ramp mechanism currently #2 in the state for climb RP%, fast scale scoring ability, reliable autonomous routines, and that super swervy maneuverable drivetrain they have come to be known for. -
67
67 made improvements to the speed of their ramp mechanism (as promised by Michael Schreiber after our week 1 analysis!) and currently have the highest climb RP % in michigan, helping them to capture 1st seed, even with 3 losses in qualifications at the Week 3 Milford event. They also have a near perfect auto RP record, and are a very consistent scale scoring robot. -
33
33 is one of the top pure scale bots in the state. The necessity of having another fully capable climbing robot to get the climb RP hurt their ranking a little, but with a speedy swerve and a multi-cube auto in the works it seemed, they have a great robot as usual! -
1918
1918, what can we say, we are still impressed with all aspects of this robot. As one of the only robots to be able to show off the triple climb capability, while still having great scale potential and auto mode. They also top the charts in Michigan for scale ownership time. -
27
27 showed off an improved buddy climb forklift this week that propels them into being top 3 for climb RP % robots in the state. They have the only perfect auto RP record, multi cube scale auto in the works it seemed combined with quick scale scoring ability. -
3357
3357 has great scale scoring ability and scale placement which is unique for FiM. Their design allows them to stack 4 cubes along the back edge of the scale which gives them a great advantage. They also have a cool two cube switch auto which gives them a good buffer at the start of the match and side hanging capability. -
3707
3707 has a very nice omni-directional drive and fast scale scoring capability. Their tethered hang also allowed them to get ~42% of the climb RP on their qualification matches, averaging 3.0 ranking points in Milford. -
5460
5460 has amazing scale scoring capability as well. Don’t believe us? Watch match 70 at Milford where they score 9 cubes on the scale by themselves. On top of this, they also have a detachable hanger that allowed them to get the hang RP in 33% of their matches. -
5561
5561 had a 50% hang RP advantage and consistent scale scoring in Milford. We are expecting to see improvements in their teleop and auton scale scoring abilities for MSC. -
3536
3536 doesn’t have as high of a hang RP% as the others, but they have had a 2 cube autonomous working since their first event in Gibraltar. The omni-directional maneuverability and scale scoring ability will put them ahead of many others for MSC. -
4362
4362 showed off their wicked fast scale scoring abilities at Milford, and their potential for a 2 cube scale autonomous. Watch out for these guys at Troy. -
245
245 had an explosive showing in Escanaba as expected. They are highly maneuverable, have great scale scoring abilities, tethered climb RP potential, and two blue banners under their belt this year. -
494
494 has a solid robot and we are expecting them to perfect their 2 cube scale autonomous mode which will give them a great advantage in qualifications and eliminations. -
4391
4391 had a great debut in Escanaba. With their omni-directional drive, scale scoring abilities, and tethered climb RP potential, we are hoping for large improvements from them in the near future for MSC finals potential. -
217
217 has great scale scoring capability, and a solid hanger that can hang with 5 seconds left in a match. We also know 217 to continue practicing and improving their robot to the point where they are a contender for Einstein, so watch out for them!
Week 4 Analysis:
Midland:
This event should be interesting because comparatively, it has a staggering number of teams (19!) that have yet to play so far this season. Since most of them are relatively new teams, it will be good to see if they have been paying attention to all of the lessons learned in the first three weeks. Eyes will be on 3770 as the only returning team with a blue banner so far this season, as the first pick on the #6 alliance from Waterford in week 2 as a capable scale bot. However, finalists in Waterford, team 5229 might have something to say about that, where they primarily worked the exchange. Finalists from Traverse City, team 5424 and finalists from Kettering#2, team 6139 are there as well. 2619 has 5 medals from the last 2 years, so don’t count these guys out. Expect a much improved scale robot with a consistent switch auto at a minimum. Overall, this event is anyone’s for the taking and it will be exciting to see who comes out on top.
Belleville:
In comparison to Midland, this event only has three teams that have yet to play. They will have to play alongside the likes of 5907 who won the Waterford event as the #6 alliance captain. They bring a nice switch auto and effective exchange cube delivery. Also 5467 who won the Southfield event as the second pick on the #1 alliance, used as an effective exchange bot will be here, as well as 6615, winners of the St. Joseph event as second pick on the #1 alliance, again was primarily playing the exchange. 2137 also enters their second event where they will be looking to improve and hopefully make a deep run into the playoffs. They were in Waterford and had a consistent switch auto, look for them to continue earning those RP and make their way to the top. Don’t count out 2834 who won 4 blue banners and 2 finalist medals in the last two years. Expect much improvement from Gibraltar and don’t be surprised if you see these guys in the finals. In addition, 1189 who was rank #2 at center Line, getting there with a nice switch auto and as a speedy scale bot. Finally, if I was into betting on high school robotics competitions, my money would be on 4405, who was rank #2 at Gibraltar as a speedy scale bot to take the gold.
Lansing
This should be a fun event to watch, as 5 teams competing here this weekend already have a blue banner on their wall. 302 decides to finally come back to Michigan and play after traveling down to Ohio to win the Miami Valley regional in week 1 as a quick and effective scale bot with fast single climb capability. 3357 is our personal favorite, who also is returning to Michigan to compete for the first time after winning at the Indiana St. Joseph event as the captain of the #1 seed alliance with a rare multi cube switch auto. Look for this highly effective scale bot with quick single side-climb to go deep into the afternoon. 494, who was rank 3 and the #1 overall pick won Kettering #2 making quick work of the scale and switch with 2 cube auto potential should do extremely good as well. Don’t forget about 5674 who was rank 1 and finalist at Waterford and 3655 who was the second pick of the #1 alliance at Gibraltar and came home with a blue banner as well. Meanwhile, 14 teams will be opening the bag for the first time at this event.
West Michigan
As one of the oldest Michigan events, this is always a good event to go to and it looks like it won’t disappoint this year either. With 4003 entering the second event after coming off an impressive rank 1 and win at Kettering#2 and the highest scale ownership time in the state, mainly due to an impressive close and far scale auto. They will be looking to get out in front again here. However 3098, who also ranked #1 and won at Center Line will be looking to take the gold for themselves (maybe together?). The other blue banner team here is 74 who was the second pick of the #1 alliance at the Indiana St. Joseph event. Other robots to keep an eye on include 85 and 3546 who ranked 5th and 6th at St. Joseph who both had effective switch autos, and 85 was an effective scale bot while 3546 was an primarily an effective exchange bot. Finalists at St. Joe, 3620 should have a working near and far scale auto. They have a history of great robot performance, so don’t be surprised to see them playing deep into Saturday afternoon either.
TMR
