Kung Fu 32: the 32 best robots of 2018

Greetings young grasshoppers! I am back!

I have awoken from my long meditation to tell you about the 32 best teams in FRC 2018: the Kung Fu 32. What does “best” mean, you may ask. Best means most likely to win Einstein. My list is centered on this idea. Thus the Dragonmaster (#1) is the most likely to win Einstein. An important aspect of the Kung Fu 32 is to acknowledge both a team’s greatest strengths and their weaknesses. In that vein, all teasing that results from this thread comes from the bottom of my heart. I have the utmost respect for the teams on this list.

Due to the nature of #2champz I have decided to make the Kung Fu 32 split evenly between each competition: 16 from Houston and 16 from Detroit. That being said, I have created the Kung Fu 32 independent of divisions; if divisions were actually completely random cough FMS cough this would be the list of teams that I think are most likely to win Einstein, in order.

When looking for teams that are the best I looked at a team’s ability to seed first then looked at a team’s ability to win eliminations. This powers down to reliability, robot performance, driver performance, autonomous, and strategy. Below I have outlined my system for quantitative ranking metrics this year.

This was a dark year for quantitative analysis of FRC teams. Because each match effectively has a set number of points that are distributed to both alliances, even component OPR this year is heavily dependant on the teams on both sides of the glass. This is different from previous years, where OPR was able to provide a much better estimate for the firepower of a specific team. It is for this reason that I have switched my primary ranking metric to Elo. Elo has gotten a lot of news this year in the FRC analysis community as a system that takes in relatively little information about the teams involved in a match and is able to output relatively accurate results. In the spirit of last year’s post, I have created my own Elo system and have run it on data starting in 2007 and up to the end of week seven. My Elo system is centered around 1500 and provides a good starting point when determining the top 32 teams in the world this year. However, it does not tell the whole story when it comes to performance this year. Teams that dominate their districts will have artificially inflated Elos just because they play so many more matches than their peers in the regional system. For instance, in the MAR district, the top 3 teams by elo (2590, 225, 25) had a combined record of 197-26. This significantly increased the Elo values for those 3 teams when compared to a team like 254, who has played exactly 30 matches this year. Elo is also unable to tell any information about climbing, which is vital to seeding this year. It is for those reasons that I am also using a significant amount of judgement in my rankings after Elo separates the ninjas from the samurai.

To prove my worth to you, I have compiled a recap of last year’s analysis which shows my predictive power.

2017 Recap
Last year’s Kung Fu 32 went on to do some extraordinary things. Our previous Dragonmaster, 254, won the Festival of Champions. The other robot in the Guru’s Few Two, 195, went on to win IRI. 1678 and 118, numbers 3 and 4 on our previous list, were finalists on Einstein at Houston. In fact, 9 of the 32 teams listed last year had Einstein appearances, and that number gets much more interesting if you consider that 13 of the 23 teams that didn’t make Einstein were eliminated by an alliance that contained another member of the Kung Fu 32. The median seed of Kung Fu 32 teams in their division was 3, the average seed of Kung Fu 32 teams was 6.4, the number of Kung Fu 32 teams outside of the top eight in their division was six, and the number of Kung Fu 32 teams that ranked first in their division was also six. This means that 50% of top seeds in divisions last year appeared on this list. Finally, every last one of the Kung Fu 32 last year made it into playoffs in their division as either the captain or the first pick of their alliance.

If you disagree with how I ranked your team, there is only one proper reaction: go win Einstein.

That’s what 973 did.

And remember, don’t try to be like Jackie Chan. There is only one Jackie Chan. Study robotics instead.

Key: Rank - Team - Elo - Elo Z-score (How many standard deviations is this team’s Elo above the mean?)

Kung Fu 32
32. 610 - Elo: 2285 - Elo z-score: 3.57
To start of this year’s Kung Fu 32, there is 610. 610 is the only shooter robot on this list, but they can go cube for cube with the best of them. Unfortunately shooter bots do have their downside, sometimes they miss. But their intake is so good, some might call it too good. 610 has the potential to seed well in their division, especially with the addition of a buddy bar, they can go for the 4 RPs every match. 610 may struggle to land the cubes when the piles start to stack up on the scale, but their outstanding exchange ability will boost their ability to play any role on a division elimination alliance.

  1. 842 - Elo: 2097 - Elo z-score: 2.7
    842 came out strong this year, rebounding from a banner-less season last year, picking up an early win at Arizona North with 254. They swooped into Las Vegas for another win, cementing them as a top 32 team. 842 showed they can play the scale game better than most. 842 has shown their over the elevator collector can give them the speed necessary to make a 4 cube autonomous mode, yet to be seen on the field. Let’s see if Falcon Robotics can snatch up another banner.

  2. 971 - Elo: 2162 - Elo z-score: 3.0
    971 built one of the few scale robots that isn’t an elevator and isn’t a double reverse 4-bar. They built a double jointed arm. The design of the arm allows the team to drive around at full height and not tip - a rare thing in this game. They field a partner climb mechanism, although it does not always work in time. By Houston I expect them to be able to place cubes neatly on the scale to allow for very high stacks. 971 needs to work on autonomous modes and not dropping cubes, but if they get those figured out then they will be a force to be reckoned with.

  3. 4003 - Elo: 2316 - Elo z-score: 3.71
    With 5 banners, currently tied for the most in the world, 4003 has dominated in Michigan this year. They won both district events they attended, Ford division, and brought home the Gold at MSC. 4003 has an advantage over many teams with an arm that can drop cubes neatly on the scale and can reach back to fill up the scale plates before going to higher levels. 4003 also has a solid intake and they can play the vault and switch, allowing them to be flexible in game play. They may not be the fastest, but if you don’t drop cubes and you are consistent, you can win the game.

  4. 1425 - Elo: 2287 - z: 3.58
    This team is the master of the strategy where you play levitate and then still lift three robots into the air. Beyond that error, 1425 was able to fight their way to the number two seed at the PNW District Championships, using their exceptional climber to secure the ranking points required. I expect this team to pull the same trick and seed very well in their division at Houston.

  5. 179 - Elo: 2120 - z: 2.81
    Team 179’s robot is like an onion; when you’re about to face them and look at it, you cry. The Children of the Swamp have shown that South Florida is their swamp, but will they do the same with Houston? They’ve demonstrated one of the most consistent partner climbs in the South, getting the climb RP in 6 out of the 10 quals matches they played at the South Florida regional. They’ve also demonstrated a 2 cube switch auton from the side and a consistent switch + scale auton in various configurations. Their far-side and near-side scale autons still need some work, but given their last event was 2 weeks ago, I’m sure they’ve fixed it. All I know is that their opponents are about to get Shrekt.

  6. 3357 - Elo: 2319 - z: 3.73
    They say a meteor killed the dinosaurs, but they are wrong. In fact it was a series of comets. The previous statement would be 100% true if I had said FiM rather than dinosaurs. 3357 has destroyed the Michigan district this year, with three event wins, a division win, and a chairman’s award at Lansing. They are currently 67-7, which is what I would call a HOT streak. With very fast and accurate scale placement and a side climb, this team is destined for an eccentric orbit around Einstein.

  7. 25 - Elo: 2389 - z: 4.05
    This team does their auto modes in a snap](youtube.com/watch?v=NHtAVhcYRS8&feature=youtu.be). But seriously, 25 has become a scale master. Team 25 is one of the few teams that has shown that a non-roller intake can work. In MAR, they were the odd man out of the power trio, getting knocked out in the finals at DCMP, but I am confident they will come back strong for Detroit. Their elevator may look floppy, but they can consistently place cubes in the scale like the best of them.

  8. 2910 - Elo: 2263 - z: 3.47
    Jester Jack would be proud of this team’s jolly juking skills. Seriously, this team has broken quite a few ankles. Their speed and precision has allowed them to single-handedly build some of the highest cube stacks I’ve seen on the scale. Despite their all-star drive team, they’ve gotten stuck in their box](youtube.com/watch?time_continue=31&v=SGiQOw2pb-4&t=0m30s) at least once. They’ve demonstrated an inconsistent 2 cube near-side scale auton and a 2 cube (attempted) far-side scale auton. The two biggest problems 2910 needs to fix for Houston is inconsistency and auton. Once those are fixed, I could see this team outscoring some of the higher up teams on this Kung Fu 32.

  9. 1538 - Elo: 2234 - z: 3.34
    When dairy meets citrus, you know something’s gotta curdle, and I think 1323 prefers their ribeye holy roasted and served on a SILVER platter. Ok, so they lost… but 1538 has built a very effective and super fast robot. At CVR, 1538 placed well and had a well thought out strategy. They look as though they might struggle with very high stacks of cubes, but they are versatile enough to play the switch. Their partner compatible single climb will go a long way in getting them into the top 8. Sometimes their hook was more like a #teamtether, which does not work to well this year… I expect 1538 go far in any division.

  10. 1241 - Elo: 2179 - z: 3.16
    Theory has it that if you pick 1114 you will win the event, but real life says otherwise. 1241 played very well recently, seeding first at ONT DCMP in the Technology division. 1241 can place quickly, has a fast elevator and good collection, but they needed to bring it all together. 1241 also has a compatible partner climb - part of the recipe for success. 1241 is also sporting 4 blue banners: 2 Chairman’s and 2 Wins, a very impressive feat. I would say that they have a chance at HoF at Detroit (but predicting chairman’s winners is not my art).

  11. 5406 - Elo: 1977 - z: 2.15
    5406 is particularly good at X-ing out their competition, even when that competition includes the likes of 1114 and 1241. Their cube handling and placement on the scale is one of the best, allowing them to quickly and efficiently place cubes, even when the scale is tipped against them. They’ve demonstrated an inconsistent 3 cube near-side scale auton and an inconsistent 2 cube far-side scale auton. Their lack of a partner climb might prevent them from seeding as high as they’d like, but they’re sure to be a first pick. Despite their low elo rating, they demonstrated that they’re deserving of this spot on the Kung Fu 32.

  12. 319 - Elo: 2372 - z: 3.97
    Big Bad Bob is certainly one of those things, and it isn’t Bob. In fact, this team’s confidence is so big that they released their CAD at the start of build season. They added their forklifts relatively late in the season, but when they did add it they faced the boss in eight out of ten quals matches. 319’s climber was so effective that even 195 instructed them to climb going into elims at NECMP. Their scale placement isn’t shabby either, and It is for those reasons that I expect big things from team 319.

  13. 2337 - Elo: 2384 - z: 4.03
    2337 engineered a great solution to Power Up. They can go for a 3 cube autonomous mode with their over the top collector/ejector. They are also quick on the placement. I think 2337 learned their lesson at MSC and will come well prepared to Detroit.

  14. 1577 - Elo: 2212 - z: 3.23
    Closing in on the Mean 16, there is Steampunk, a team that, to put it simply, chose the wrong year to be good (if the name has anything to do with it). They’ve gone through eliminations undefeated at every event they’ve been to. At the Israel DCMP, they got the climb RP in 7 out of the 11 quals matches they played. Their autons are steamjunk, though: at the same event, they showed a 2 cube switch auton and an inconsistent 2 cube scale auton. They’ve obviously got a lot to do in the auton department, but 1577 clearly has the climb capacity to seed high and cube competence to win matches.

  15. 217 - Elo: 2328 - z: 3.77
    [They were caught
    In the middle of the leaderboard
    They looked round
    And they knew there was no turning back
    Their mind raced
    And they thought what could they do
    And they knew
    There was no help, no help from climbs
    Sound of the scale
    Beating in their vault
    The thunder of cubes
    Tore them apart
    You’ve been


The Mean 16
16. 2471 - Elo: 2187 - z: 3.12
What better way to start the Mean 16 than with the Mean Machine! 2471 has shown time and time again that they are on par with the best of the best. They’ve demonstrated a consistent 2-cube far-side scale auton and a 3-cube near-side scale auton in competition, and they’ve showed off some even more impressive scale autons for Houston. However, their ability to climb with a partner fell short of expectation. Their iron cross platform was not very consistent at their last event; they got the climb RP in 4 of their 12 quals matches. There is definitely room for improvement, but if anyone’s up to the job, it’s the Mean Machine.

  1. 2767 - Elo: 2365 - z: 3.94
    Rode up the rankings
    Broke the limit, they hit the top
    Went through to MSC, yeah MSC, and they had some fun
    They met some chickens
    Some chickens who gave a good time
    Swerved round the auton
    Played all the fools
    Yeah yeah they, they, they blew their minds
    And the divisions were shaking at the knees
    Could they come again please
    Yeah them chickens were too fast
    You’ve been


I was thunderstryked when they lost.

  1. 225 - Elo: 2356 - z: 3.9
    TechFire brought the heat to MAR this year. Their elevator and arm combination allow for amazing scale placement, even at high levels of play. They have also been able to show off a 3 cube scale autonomous at MAR DCMP. This allowed them to play on the number one seeded alliance and win the event, which leaves them to enter champs with a record of 65-9! TechFire does need to bring down the heat in the null zone sometimes, but that won’t stop them from scorching their division.

13 1325 - Elo: 2393 - z: 4.07
This team is easily one of the best to come out of Canada this year. With five banners, a DCMP win, and a tie for first place in points in the Ontario District, this team is looking to offset the myth that there are only 4 teams in Canada that are worth talking about. Dominance on the scale isn’t everything, however, and I expect 1325 to suffer from their lack of a buddy climb in the rankings going into Detroit. That being said, 1325 is sure to be a first pick in their division and I have every reason to believe that they can go all the way.

  1. 1619 - Elo: 2131 - z: 2.86
    At first glance, this robot is entirely unique as the only member of the Kung Fu 32 with a horizontal intake. However, even they will admit that the design is nothing that we haven’t seen before. And boy is that intake effective. Well, it’s generally slightly more effective than that; just watch any Idaho match. 1619 was able to seed above 148 and 118 at Colorado, becoming one of the few teams in history to break up the Space Cowboys. 1619 will be able to seed high due to the effectiveness of their partner climb, and the over-the-top placement allows them to easily grab and place fence cubes, potentially leading to a 3+ cube auto. That being said, I have not yet seen any auto mode from them that places more than 2 cubes.

  2. 2481 - Elo: 2491 - z: 4.52
    Although 2481 can not sail to first seed without a climber, there is no doubt that they will dominate the scale game in any division. At week 4, 2481 was playing like many week 6 robots. Now, after many weeks, I am excited to see what they will bring to Detroit: buddy climber? Better auto modes? More driver practice? All of it? Their swerve drive allowed them to easily nail down more complex auto routines, and their rotating arm allows them to place and pickup on either side of their robot with little trouble. I don’t see a combination like this very much, but when I do it devastates.

  3. 3310 - Elo: 2291 - z: 3.6
    We’ve got a Black Hawk down… we’ve got a Black Hawk down! 3310 has had their issues, but they are an amazingly quick robot that can pass cubes through their elevator when necessary and they can drift around the field like no one else (except maybe 7179…). 3310 was ejecting 3 cubes onto the scale in autonomous, but did not get every cube on, but I have high hopes for their autonomous modes in Houston. 3310 boasts a very effective intake and elevator, both of which will be key to getting to Einstein.

  4. 2590 - Elo: 2558 - z: 4.83
    2590 brought the Fury to MAR. They decimated the competition with a regular season final record of 69-6 and earned 3 blue banners so far. Their consistent double partner climb has allowed them to seed first at every event they have attended so far. Between their autonomous and teleop scale ability they have been able to win MAR DCMP. It will be interesting to see them play at a higher level of competition at Detroit, but they will go far.

The Great 8
8. 148 - Elo: 2253 - z: 3.42
To start off the Great 8, there is the Robowranglers, the only team that started with a black belt. This team didn’t wrangle their opponents as well as they wanted to last year, so this year they’ve come ready to wrangle everyone, even tall alliance partners. If that wasn’t enough, they’ve also demonstrated the fastest, smoothest elevator I’ve seen in any robot, and a touch-it-own-it collection that works most of the time, with the exception of occasional auton mishaps. On the subject of auton, they’ve demonstrated a 2.5 cube near-side scale auton and a consistent switch + scale auton in various configurations, but their far-side scale auton lacked tuning. This is fine; I fully expect them to have tuned, consistent, and more capable auton modes ready for Houston.

  1. 2122 - Elo: 2313 - z: 3.7
    2122 slowly ramped up their season this year. But their robot drastically changed from Arizona north to Idaho. Climbing: different. Intake: different. And the changes paid off. After an unfortunate outing at Arizona North, 2122 picked up their game](youtu.be/UuiRc2g4kC4?t=2m11s) at Idaho and Canadian Rockies; getting back to back wins. Team tators used consistent multi-cube autonomous modes throughout Idaho and Canadian Rockies. This will help them gain control of the scale early, but more teams can do 3 cube autonomous modes, so they will need to speed up if they plan to win.

  2. 118 - Elo: 2478 - z: 4.46
    Despite their stacked division, the Robonauts have shown that they’re roboNOT going out without a fight. At the Colorado regional, they were not disillusioned when 1619 chose 148 over them, and went on to beat them in the finals. While some of their auton modes could use some tuning, they’ve demonstrated a consistent 2 cube far-side scale auton and a 2 cube plus missed cube near-side scale auton. I am expecting a 3+ cube autonomous at Houston. Their cube placement, second to none, makes up for their lackluster autons. Not only can they place the most common cube configuration in the most optimal way on the scale, they can place high … really high. We could definitely be seeing another 118-1678 alliance on Einstein.

  3. 2056 - Elo: 2600 - z: 5.03
    OP robotics has begun a new streak this year. A streak of breaking streaks. In 2016, they lost their first regional. In 2017, they lost their first IRI since 2012. This year, they had their very first bannerless non-championship event at Waterloo. This is not to say that they have not built a beautiful robot this year, as they have still taken home 4 banners, won the Ontario district, and have the second highest Elo in the world. Their partner lifting forks are devastatingly effective, and they are the only team to accomplish a 3 cube auton mode without somehow passing the cube through or over their robot.

Highscore 4
4. 1323 - Elo: 2220 - z: 3.27
MadTown Robotics is one of the best scale robots: fast, tall, and well practiced. 1323 has used their swerve drive for some great plays including keeping 1538 away from fence cubes, and dodging 1678. But beyond their drive train, 1323 boasts a very effective 3 cube switch autonomous with time to spare. 1323 can play the switch, the scale and climb like no other. Together this combination will be devastating at worlds and provide for some epic matches. Already, at CVR, 1323 played against 1538 and 1678 (in different matches), winning both. I fully expect 1323 to make it to Einstein this year.

  1. 1678 - Elo: 2443 - z: 4.3
    The Citrus Circuits have been to Einstein 5 years in a row, and they’re looking to keep their streak alive. 1678 released some early 3 cube auton modes, and since then I knew they would leave a sour taste in the mouths of their opponents. Their platform is grate; 1678 has one of the most consistent partner climbs in the world. At their last event, the Central Valley Regional, they only missed 2 climbs during quals, and at the Sacramento Regional, they missed none. 1678 always loves to help out other teams, and that’s why they are putting together some tips and tricks.

Guru’s Few 2
2. 195 - Elo: 2601 - z: 5.03
The CyberKnights are no stranger to the Guru’s Few Two. With a season record of 51-3 and a 83% record on 4 RP matches at their most recent event, 195 took a look at the New England district and said Good Knight. They are 3 for 3 on event wins, and they have the highest Elo in the world. They have a very consistent double cube auton where they go for 3 and miss the second and can even be seen starting to go for a 3.5 cube auton in the elims at NECMP, which leads me to believe that the next time I see them compete they will have a coveted 4 cubes coming out of autonomous mode by Detroit. They have the most consistent climber in the North, and their toothpick style elevator takes this team to insane heights. It is for these reasons that I welcome 195 back into the Guru’s Few Two.

The Dragonmaster

  1. 254 - Elo: 2556 - z: 4.82
    254 has this game on Lockdown. The Cheesy Poofs have been a force to be reckoned with since their very first match. Sounds familiar? This year’s Dragonmaster has demonstrated they can play every aspect of Power Up, from vault to scale, at the highest level. They have had a perfect season, winning every match they’ve played in. At their last event, they became the only team to achieve a 4.00 ranking score after qualification rounds. They’ve demonstrated 3.5 cube autons for the near-side scale, the far-side scale, and the switch. It is almost a given that 254 is JUST GONNA SEND IT to Einstein this year.

Great analysis but it appears Master Chan is yet to master the art of link formatting :wink:

Thanks for including us in the Kung Fu 32! Just a small correction:

"Their lack of a partner climb mightprevent them from seeding as high as they’d like, but they’re sureto be a firstpick."

I will cry if 1323 doesn’t finally make Einstein. Their appearance is past due.

knocks on wood

Always well done! Looking forward too see how champs pan out.

Top 16 Houston List

  1. 254 (hopper)
  2. 1678 (newton)
  3. 1323 (roebling)
  4. 118 (netwon)
  5. 2122 (turing)
  6. 148 (hopper)
  7. 3310 (newton)
  8. 1619 (newton)
  9. 2471 (turing)
  10. 1577 (galileo)
  11. 1538 (turing)
  12. 2910 (carver)
  13. 179 (roebling)
  14. 1425 (carver)
  15. 842 (turing)
  16. 971 (newton)

3. 1678
6. 118
10. 3310
12. 1619
31. 971

7. 2122
16. 2471
23. 1538
30. 842


  1. 254
  2. 148

4. 1323
27. 179

24. 2910
28. 1425

18. 1577


It appears that there were some hidden characters in my links… almost as sneaky as 5406’s buddy climb! The problem is resolved now.

1325 has also nailed the three cube auto without passing through or over - here is a fun match where they match 2056.

Of the many deserving Ontario teams, 1114 is worth consideration for cubing speed that matches 1241, 2056 and 5406, a three cube auto and a hang that is compatible with many of their partners. 4678 also merits recognition for their three cube auto, expert placement and reach over multiple cubes.

Detroit division breakdowns

13. 1325 - Elo: 2393 - z: 4.07
29. 4003 - Elo: 2316 - z: 3.71

14. 225 - Elo: 2356 - z: 3.9
17. 217 - Elo: 2328 - z: 3.77
25. 25 - Elo: 2389 - z: 4.05
26. 3357 - Elo: 2319 - z: 3.73

2. 195 - Elo: 2601 - z: 5.03

15. 2767 - Elo: 2365 - z: 3.94
32. 610 - Elo: 2285 - z: 3.57

11. 2481 - Elo: 2491 - z: 4.52
19. 2337 - Elo: 2384 - z: 4.03
21. 5406 - Elo: 1977 - z: 2.15

5. 2056 - Elo: 2600 - z: 5.03
9. 2590 - Elo: 2558 - z: 4.83
20. 319 - Elo: 2372 - z: 3.97
22. 1241 - Elo: 2179 - z: 3.16

Whats with my luck and hard divisions… first dow msc, then daly, then dte msc, now we here

Kung Fu Results:

Won Finals:

  1. 254 (hopper)
  2. 148 (hopper)

Round Robin:
3. 1678 (newton)
4. 1323 (roebling)
12. 1619 (newton)
24. 2910 (carver)

Division Finals:
6. 118 (netwon)
7. 2122 (turing)
10. 3310 (newton)
27. 179 (roebling)

Division SF:
18. 1577 (galileo)
30. 842 (turing)

Division QF:
16. 2471 (turing)
23. 1538 (turing)
28. 1425 (carver)
31. 971 (newton)

Welcome back, grasshoppers!

I have returned to give a quick recap of the Kung Fu 32’s performance now that both championships have wrapped up.

First, let’s see where everybody ended up:

Einstein Winners:
1	254
8	148
15	2767

Einstein Finalists:
17	217
24	2910
26	3357

Einstein Semifinalists:
2	195
3	1678
4	1323
5	2056
12	1619
22	1241
29	4003

Division Finalists:
6	118
7	2122
9	2590
10	3310
27	179

Division Semifinalists:
11	2481
13	1325
14	225
20	319
21	5406
25	25
32	610
18	1577
30	842

Division Quarter-finalists:
16	2471
19	2337
23	1538
28	1425
31	971

I also have made scatter plots relating the Kung Fu 32 rankings to both district points (as calculated by AriMB) and rank on the final FRC Top 25:

Good luck to everybody in the off season and I can’t wait to see which teams make the list next year!

(post withdrawn by author, will be automatically deleted in 24 hours unless flagged)

We came all this way for a withdrawn post


I was wondering if there would be a 2019 one but then I realized that these used to be posted before champs

Oh. Good day then

He graduated last season, so I’m guessing he might be taking a break

1 Like

Ah. Never thought he was a student due to how difficult it seemed to make these lists.