World Qualification Ranks

[BrennanB]

Thread/Google doc (https://docs.google.com/spreadsheets/d/1F29pXEYzUWdkIaHhWW0lsUKNW4p8kJPje_AWSE0ckEM/edit#gid=0) that is updated erratically. It's pretty simple, just merges all of the qualification data into one place.

Only done for fun, not meant to be an actual ranking list of best to worst robots, it's just a fun thing to look at.

Current Top 25

Rank Team Qual Ave
1 1114 141.5
2 148 116.1
3 2481 113.28
4 987 106.3
5 67 96.33
6 3130 95.66
7 2451 94.57
8 525 92.22
9 1519 90.41
10 744 88.9
11 4488 88.58
12 4048 84.66
13 3620 84
14 1706 83.85
15 118 83.8
16 176 82.55
17 1025 82.37
18 4330 80.5
19 1403 80.33
20 2386 79.33
21 1209 79.28
22 2607 78.91
23 5172 77.66
24 3824 77.33
25 1208 77.28

WEEK 1

Rank Team Qual Ave
1 148 116.1
2 987 106.3
3 3130 95.66
4 525 92.22
5 1519 90.41
6 744 88.9
7 4488 88.58
8 118 83.8
9 1403 80.33
10 2607 78.91
11 5172 77.66
12 3824 77.33
13 1024 75.5
14 4623 75.22
15 1983 75.08
16 1640 73.33
17 1477 73.3
18 2342 72
19 4859 70.55
20 179 70.4
21 1592 69.8
22 4539 69.55
23 348 69.2
24 4624 68.55
25 3242 68.3

WEEK 2 (In Progress)

1 1114 141.5
2 2481 113.28
3 67 96.33
4 2451 94.57
5 4048 84.66
6 3620 84
7 1706 83.85
8 176 82.55
9 1025 82.37
10 4330 80.5
11 2386 79.33
12 1209 79.28
13 1208 77.28
14 2609 76.28
15 857 71.42
16 931 71.33
17 3612 68.85
18 4256 68.42
19 5053 68
20 1732 67.71
21 4522 67.71
22 610 67.62
23 4779 67.44
24 16 67.42
25 5478 67.37

[dodar]

47.8 difference from #1 to #25; that is amazing.

[zachrobo1]

I ran some general stats:

Mean: 43.238
Median: 41.2
Standard Deviation: 13.624
Range: 119.28
Minimum: 22.22
Maximum: 141.5
Count: 955

[BrennanB]

With week 3 events started/starting here are updated ranks (https://docs.google.com/spreadsheets/d/1F29pXEYzUWdkIaHhWW0lsUKNW4p8kJPje_AWSE0ckEM/edit?usp=sharing)

Worldwide

1 1114 142.9
2 2481 117.8
3 148 116.1
4 1678 109.6
5 987 106.3
6 2451 104.8
7 254 104.4
8 67 97.41
9 3130 95.66
10 525 92.22
11 1519 90.41
12 744 88.9
13 4488 88.58
14 1025 88.08
15 1208 86.4
16 118 83.8
17 2386 81.8
18 1403 80.33
19 4048 79.83
20 2607 78.91
21 4330 78.4
22 2137 77.83
23 5172 77.66
24 1706 77.5
25 4522 77.5

Week 2 Only

1 1114 142.9
2 2481 117.8
3 1678 109.6
4 2451 104.8
5 254 104.4
6 67 97.41
7 1025 88.08
8 1208 86.4
9 2386 81.8
10 4048 79.83
11 4330 78.4
12 2137 77.83
13 1706 77.5
14 4522 77.5
15 3616 77
16 314 76.83
17 3612 75.4
18 176 74.83
19 217 74.33
20 3620 74.16
21 1986 74.1
22 5053 73.83
23 701 73.8
24 379 73.66
25 4946 73.6

[Chief Hedgehog]

That is an incredible list - and I love seeing 3130 in the top 10!

I was able to watch them at Northern Lights and they are incredible! The Errors have had great robots for the last three years that 4607 has been around and we always look to them for great ideas! We actually adopted some of their ideas from the Week Zero event (hosted by 2472 and 2052)...

My goodness I love this time of year!

[faust1706]

Your list doesn't filter out teams that aren't qualified for worlds (I think). I don't believe 1706 is qualified for worlds (hopefully not yet at least).

[BrennanB]

Your list doesn't filter out teams that aren't qualified for worlds (I think). I don't believe 1706 is qualified for worlds (hopefully not yet at least).

It's not actually supposed to filter out anyone.

[Galum]

Any chance this would be updated with week 3 events? Pretty please with sugar on top? :)

[AWoL]

I would love to see this updated for week three as our QA this week was 98.58 :D

[theawesome1730]

Our qual average was 126.3 or something like that placing us as number 2 if we had competed in week 2. Eager to see who's moved around

[BrennanB]

Updated as all the events just finished tonight :)

Both median and average jumped by 3 points, probably due to some teams starting to compete in their second event/people watching matches and learning how to play the game.

Of the top 100 Co-op per match scores came from the following weeks:

Week 1: 29%
Week 2: 37%
Week 3: 34%

Co-op scores in the top 100 are broken down here:

Max: 440 (Co-op stack in 11/12 matches) (Week 2)
Average: 321.2 (Co-op stack in 8.03/11.67 matches)
Min: 280 (Co-op stack in 7/12 matches) (Week 3)

And your most co-opiest teams are:

4539 94.4%
5782 91.7%
1649 90.0%
348 90.0%
2386 85.0%
1519 83.3%
78 83.3%
624 80.0%
2481 80.0%
744 80.0%
456 80.0%
4471 80.0%
348 79.2%
176 79.2%
4906 79.2%
178 79.2%
3130 77.8%
4859 77.8%
4564 75.0%
4048 75.0%
1024 75.0%
217 75.0%
4381 75.0%
133 75.0%
2079 75.0%


Litter however has been on the up and up, and teams have been using it more now than before.

Of the top 100 Litter score per match came from the following weeks:

Week 1: 14%
Week 2: 39%
Week 3: 47%

Seems like more and more litter is getting put in cans.

Finally the rankings: (https://docs.google.com/spreadsheets/d/1F29pXEYzUWdkIaHhWW0lsUKNW4p8kJPje_AWSE0ckEM/edit#gid=0)

Worldwide:

1 1114 142.9
2 1519 136.08
3 1730 126.3
4 2056 124.5
5 624 124
6 1983 121
7 2481 117.8
8 148 116.1
9 1619 115.3
10 1023 115.25
11 1678 109.6
12 987 106.3
13 2451 104.8
14 254 104.4
15 1523 102.7
16 3663 102.41
17 2974 101.3
18 234 100.33
19 225 98.58
20 2122 98.5
21 2996 97.9
22 67 97.41
23 3230 97.4
24 744 97.1
25 1501 96.91


Week 3:

1 1519 136.08
2 1730 126.3
3 2056 124.5
4 624 124
5 1983 121
6 1619 115.3
7 1023 115.25
8 1523 102.7
9 3663 102.41
10 2974 101.3
11 234 100.33
12 225 98.58
13 2122 98.5
14 2996 97.9
15 3230 97.4
16 744 97.1
17 1501 96.91
18 1806 96.8
19 4451 96.7
20 135 95.75
21 1720 95.41
22 2852 95.3
23 662 95.1
24 192 94.6
25 246 94.33

[tindleroot]

Woo! 4 highest seeds from Kokomo are all on the week 3 list, with the top 2 being in the overall list!:D

[wireties]

Those coop points really help! We had a weird weekend, seeded first and won Alamo with zero coop points all weekend - arghhh.

[BobbyVanNess]

I don't mean to take away from the teams on these lists, but qualification scores are highly dependent on the depth of the particular event, and I think that it misrepresents the abilities of teams. Personally, I think it'd be more valuable to compare teams on a metric that better identifies their individual performance, like OPR.

That being said, the well deserving are rising to the tops of these lists anyway.

[Nathan Streeter]

I don't mean to take away from the teams on these lists, but qualification scores are highly dependent on the depth of the particular event, and I think that it misrepresents the abilities of teams. Personally, I think it'd be more valuable to compare teams on a metric that better identifies their individual performance, like OPR.

That being said, the well deserving are rising to the tops of these lists anyway.

Agreed that depth of particular event will impact these (although so far less than 2014 that it really isn't funny).

Unfortunately, to my knowledge, there's no way to determine how much was scored in each category in each match this year... the twitter feed used to supply this data, enabling 'category OPR.' Without that, this is the best that can be done without a ton of effort.

(Edit: just realized I thought this was in Ether's thread with the category averages... guess the second bit is pretty much irrelevant. Ether's thread with top 25 in each average of category is quite interesting; look out for Ed Law's week 3 update for his OPR/CCWM spreadsheet)

[Ether]

Unfortunately, to my knowledge, there's no way to determine how much was scored in each category in each match this year... the twitter feed used to supply this data,

Correct.

enabling 'category OPR.'

"Category" or as some call it "Component" OPR is still possible, by using the category (component) scores in the Team Rankings table. Problem is, DQs muddy up the waters because the FRC API that is supposed to provide DQ data is broken.

Without that, this is the best that can be done without a ton of effort.

That's where Ed Law comes in :-) He devotes a lot of effort to make the most out of the data that is available.

[BrennanB]

I don't mean to take away from the teams on these lists, but qualification scores are highly dependent on the depth of the particular event, and I think that it misrepresents the abilities of teams. Personally, I think it'd be more valuable to compare teams on a metric that better identifies their individual performance, like OPR.

That being said, the well deserving are rising to the tops of these lists anyway.

Exactly. OPR is a better guess at what individually the teams can do than qual average scores. And obviously watching their matches is an even better guess as to how they will preform in their next match.

What makes the list interesting is how weird seeding is this year. It's by far more indicative of a team's strength than previous years with win/losses and the list is only a data source for looking at how seeding changes.

The rankings themselves are no more useful than people keeping track of world high scores or many of the other things that we look at to an extent. Lists made of averages, OPR, or CCWM ultimately are for fun. You can't watch every single match to know who is better than who, but you can give educated guesses.

This data however can be used to quantify game trends to some extent, like the massive increase in litter in cans from week 1 to 2, and 2 to 3 to a lesser extent. More interesting things are in the data, you just have to look.

[GuyM142]

Is it possible to do a similar ranking for finals averages?

[Addison4300]

It says my team was at Orlando? We were at week1 Dallas, the rest of the stats are correct however.

[cglrcng]

I don't mean to take away from the teams on these lists, but qualification scores are highly dependent on the depth of the particular event, and I think that it misrepresents the abilities of teams. Personally, I think it'd be more valuable to compare teams on a metric that better identifies their individual performance, like OPR.

That being said, the well deserving are rising to the tops of these lists anyway.

YES they certainly are...and then capping them w/ Littered RC's over and over.

How can you possibly use Q Points Average to compare a midling robot in an 8 Q Match Regional (See Week 4 Virginia), against a 13 Q Match Regional (See Week 4 Waterloo), now add 1114, 2056, & 5406 to that mix? You cannot....More chances at success for 1114 & company, more chances at failure for many, many others. Or, the opposite fewer chances to stub your toe (or even have others do it for you in blind draw qualifying...I'm talking about non-movers, not the attempters).

The whole thing is percentage based to the very last few matches. How is it, that disparity in number of Q Matches was possibly left in?

Per Regional I understand...level playing field for all competing there is OK (each plays the same amount of matches). Comparing QPA at Virginia vs. Waterloo is not OK...Even without the super magic dream team alliance partners. (Outlier(s)...LOL Ummm, different planet or solar system maybe).

Does the OPR calculation set, somehow correct for that major disparity? If so, how? Inquiring minds want to know.

[cglrcng]

Please don't take any of the below (Team 5406), stated as personal, since it shows (w/ the 7 mid 200 point + playoff matches), that you contributed to each quite well, without seeking out the videos...yet!

Just trying to prove a point.

Put any absolutely "never move dead bot" (in the right place, on or off the field of course), w/ 1114 and 2056 in any Q Match, and their (the dead ones), personal QPA will exponentially increase without ever contributing a single point (call that the QPA Inflation).

Then increase the amount of matches from 8~13, and that difference gets larger & larger every match in the QPA.

This year you cannot judge even a single bot by their QPA, unless you know personally, that it is a major points contributor, or not.

Or just look at the easily seen spreads between the top QPA Averages, then their (as built/used), specialties are easily seen in the other data and very evident, the actual qualifying scores help here), as pertains to co-op, Auto, Container, Tote, & Litter.

(I Just noticed only 1 point total Litter in 10 Q matches in Dallas 148? You were the vid release masters of the over the wall bent litter stuff, and it was ruled legal). Headed back to watch the videos again now...Stacking, capping totes gains more points, and time is valuable for sure). I know w/ 987, I'd give them all the litter they can handle...but over that many Q matches? Hmmmm...

But, midrange in the QPA Rankings...The rest of the data avail. and listed is just as much junk too. You score Litter points, Tote points, RC points (I can do nothing...I get the same points, and credit, as the real contributors). TYVM. "The eyes still have it," by watching what robots can, and do actually, do.

Your eyes is all you have this year (thank you YouTube, Live streamers, videographers, and other contributors)...And the Playoff points averages the alliances do score.

There is no tossing the Last Stand at Waterloo Playoff scores or the 284 point Match 20 among others...They are as real (and scary & inspiring, all at the same time), as the master bots (and teams), that designed, built, operate, & posted them.
________________
Waterloo 30 bots, 13 Q Matches ea., total 65 Q matches (QPA Spread 183.46~53.38). At least 2 really high scorers. Third bot, I can't tell until I see the actual videos, as the QPA inflation rate takes over somewhere in there. (No offense please to 5406 @ 135.69 QPA, you earned it). Still though QPA inflation/deflation is real.

Virgina 64 bots, 8 Q Matches ea., total 86 Q matches (QPA Spread 83.25~21.25). Average scorers. No QPA Inflation rate appearing.

Difference.....5 more Q matches to either shine in, or stub your toe.

Comparisons you can possibly make between robots at 1 event today (Week 4), and the other based on any of the provided data (except of course 2 bots we already know are absolutely fantastic) ~ NONE WHATSOEVER. (You are comparing the economies of S. Africa and it's diamond and gold mines, to Botswana).

Now the Playoff scores and percentages are absolutely real, as long as all 6 bots were on the field or in the vicinity. As well as Playoff Alliance averages.

And some amazing averages & scores they are indeed!
_______________
Same 2 high scorer QA Inflation/Deflation applies to Dallas in Week
1, though w/ the weather affecting the event like it did, it isn't fair to judge that event. Just the bots that put up the great scores.

[Sunshine]

Does anyone know the 3 highest scores ever achieved so far this year? Or scored during elims vs qualifications? I know 1114, 2056 and 2935 scored 284 in their qualifications match. Just curious how our 216 points in semi-finals ranks. Kids are telling me it's the second best in world and first in country so far. I'm finding it hard to believe.
Thanks

[Ether]

Does anyone know the 3 highest scores ever achieved so far this year? Or scored during elims vs qualifications? I know 1114, 2056 and 2935 scored 284 in their qualifications match. Just curious how our 216 points in semi-finals ranks. Kids are telling me it's the second best in world and first in country so far. I'm finding it hard to believe.
Thanks

As soon as all the Week4 qual matches are done, I'll post my Top25 (http://www.chiefdelphi.com/media/papers/3118) list again.

[Sunshine]

Thanks

[nuclearnerd]

(No offense please to 5406 @ 135.69 QPA, you earned it). Still though QPA inflation/deflation is real.

No offense taken. Our QPA was definitely higher than we contributed alone. Certainly our matches with 1114 and 2056 were our top scorers (252 @ q24 and 202 @ q54). We could put up 80 to 100 point ourselves if nothing went wrong. Our OPR of ~80 is closer to the actual mark. Still, we were able to post 190 points in q61 without 1114 or 2056, so we weren't completely carried :)

[ijonny5]

I scanned TBA this morning. Sacramento put up 226 in the semi's and 218 in the finals yesterday, so only for a brief moment did we have the highest score in the US. If you have time, you should check out the Waterloo Regional stats, 1114, 2056 and 5719's LOWEST elimination score was 200 and the rest were over 224!

It was great to bring home another regional win with 2062 and 2530 yesterday. See you in St. Louis!

[nuclearnerd]

Still, we were able to post 190 points in q61 without 1114 or 2056, so we weren't completely carried :)

Thanks to 1676 and 1305 by the way for being amazing alliance partners that match.

[Ether]

As soon as all the Week4 qual matches are done, I'll post my Top25 (http://www.chiefdelphi.com/media/papers/3118) list again.

Here's the raw data (in XLS format) for Week4 Qual Match Results and Team Rankings:

http://www.chiefdelphi.com/media/papers/3120

You can search, sort, and plot it any way you like.

[cglrcng]

No offense taken. Our QPA was definitely higher than we contributed alone. Certainly our matches with 1114 and 2056 were our top scorers (252 @ q24 and 202 @ q54). We could put up 80 to 100 point ourselves if nothing went wrong. Our OPR of ~80 is closer to the actual mark. Still, we were able to post 190 points in q61 without 1114 or 2056, so we weren't completely carried :)

I saw some video also and I believe every word you posted (there is still an inflation value with the QPA's when there are a couple to a few exceptional bots present, especially for lesser contributors than those that can put up 100+ alone on the board). I knew you were a decent to higher contributor (even before I saw the vids), as even the best can't skyrocket to 282 without help, the game is limited in time. (You put up 190, you were never carried!) Great Job!

Thank you for not taking any offense. The data Ether has provided today (Weeks 1~4), and OPR vs. Live Scouting data, and discussions about the disparity in # of Q matches at events, and between events (and he does attempt to correct for that, but given the lack of actual per team vs. per alliance real data), I don't know exactly how much trust can be put in that this year. (I do realize that he is doing the best w/ data avail. though, and the extreme work put in to deliver that is fully appreciated BTW).

Then, I actually went and looked at the OPR data, and how he calculates it. I trust it a bit more now after comparing actual game play vids (action), for just kicks watching random game play in Q matches over many events for a few hrs., then some playoff matches (of course the cream always rises to the top there), and compared my opinions only of indiv. bot gameplay vs the OPR listing locations Weeks 1~4...It seems fairly somewhat true to form, robot to robot though, given their distinct capabillities (or not), if you look at it and, add scale top to bottom~ which is what the OPR actually does.

TY to Ether for making the data avail to all here, and explaining how you do the OPR calcs a bit more clearly, also.

[c.shu]

I scanned TBA this morning. Sacramento put up 226 in the semi's and 218 in the finals yesterday, so only for a brief moment did we have the highest score in the US. If you have time, you should check out the Waterloo Regional stats, 1114, 2056 and 5719's LOWEST elimination score was 200 and the rest were over 224!

It was great to bring home another regional win with 2062 and 2530 yesterday. See you in St. Louis!

Our alliance (2137, 1918, 4967) put up 228 in Qualification match 47 at West Michigan.

[Boltman]

It says my team was at Orlando? We were at week1 Dallas, the rest of the stats are correct however.

Lots of errors in event locations had to fix 5 in my list of 29..otherwise seems accurate and a great resource. Thanks to those who compiled it.