1st Seeds Win

[Jim Zondag]

It is pretty hard to win if you are not in the top 2. The past 4 years of combined results are in the attached chart. More than 80% of the time, the winners are one of the top 2 alliances.
As Al said above, this is because the best teams seed highest. This means that the seeding system works properly.

[TRWSHSHLX]

There are a few components to upsets in regionals. Since the qualifying rounds schedule is totally random sometimes a good team might be put against another good team but with better alliance partner and lose those matches. That'd cause a good team to be ranked lower than they really should.
Secondly there are always the problems of debugging. Like some posts mentioned, some teams do not have all the bugs worked out until the elimination rounds. Those teams are usually overlooked but if picked, it can be a huge factor.
And lastly, the ranking system that FIRST uses to rank the teams are purely based on offense. The GDC encourages offense and an offensive robot or a team that's always with an offensive robot (not likely) will get really high ranked. However, there's no mention of defense in the ranking system what so ever. If a robot is a mostly defensive robot, it will most likely rank relatively low and sometimes they're overlooked but can be another major game changer because in elimination rounds, it's the point difference between alliances that matters, not as much as you can score, like in qualifying rounds.
All in all, a successful alliance will have both offense and defense. And the alliance that can execute always wins.

[nikeairmancurry]

Originally Posted by Jim Zondag

It is pretty hard to win if you are not in the top 2. The past 4 years of combined results are in the attached chart. More than 80% of the time, the winners are one of the top 2 alliances.
As Al said above, this is because the best teams seed highest. This means that the seeding system works properly.

Thats why you one of the best Jim, I always love your charts... Perfect way, to show the spread the winners of events...

[Kett Captains]

An upset just happened this weekend at the Michigan State Championship. The 8th seeded alliance, (74, 548, 3098), beat the 1st seeded alliance, (217, 469, 201). Seems like teams definately overlook performance on saturdays before elimination.

[DonRotolo]

Well, the #1 seed at Virginia also didn't win this weekend.

[mwilbur]

Originally Posted by DonRotolo

Well, the #1 seed at Virginia also didn't win this weekend.

The #1 seed at Virginia didn't win this weekend, but the BEST alliance DID. Kudos to team 25 for making great selections. We knew that we were going to have to be 10 points up going into minibot deployment with your super fast minibot on the field, but we didn't manage to pull it off. A last minute pneumatics leak on our robot right before the start of the second finals match didn't help, but it still would have been too close for comfort had that not happened. The competition ranking system is good for a ballpark idea as to which teams are doing the best, but the scouting part of it is critical to put the best alliance together.

[TEE]

Originally Posted by Kett Captains

An upset just happened this weekend at the Michigan State Championship. The 8th seeded alliance, (74, 548, 3098), beat the 1st seeded alliance, (217, 469, 201). Seems like teams definately overlook performance on saturdays before elimination.

Goes to show the intensity of MSC...

For those who didn't watch, 217, 469 and 201 would have won the second match if they got a minibot up the pole for 3rd place, and in the third match, 3098 stopped 469 from deploying (that was the part of strategy 217, 469 and 201 missed, or simply didn't execute: defending against minibots), and 469 had our alliance's fastest minibot. The tube-scoring was really close in all three matches.

On a different note, I personally was really surprised that 548 wasn't chosen sooner... I expected them to be picked by one of the first 3 or 4 alliances.

[XaulZan11]

Originally Posted by TEE

The tube-scoring was really close in all three matches.

Tube scoring was close in most of the elimination matches, which is not that surprising. My biggest complaint with the game is how being an amazing tube scoring alliance is not that big of an advantage over a good tube scoring alliance. If you score 6 logos while the opponent gets 4, you only get 12 more points. The decreasing marginal returns to tube scoring makes the game a little too easy strategically; no longer are you having the tough decision to place one more tube or go to endgame, like you did in 2007 (which also had spoilers).

This was really evident in the elminations at MSC. Since both alliances could consistently and easily score 3-4 logos, most matches came down to autonomous and minibots. If you had a significant disadvantage in auto (like 3-1), you needed to get first and second place in the minibot race to make up those 24 points. 469 just picked a really bad time to miss their auto and minibot in the final 2 quarterfinal matches (I believe they were 12/13 on each before missing both in back to back matches). In my opinion, that was the deciding factor. When the top pick cannot do the two most important things, the 8 seed has a chance.

[DonRotolo]

Originally Posted by mwilbur

The competition ranking system is good for a ballpark idea as to which teams are doing the best, but the scouting part of it is critical to put the best alliance together.

Bingo: That's the point of the thread. Perhaps the ranking system is not a perfect indicator (and never will be), but as you said, it's not too far off.

[jckwik]

The ranking system, when a win-loss record, does not do justice to a teams actual ability, and occasionally will be very wrong. For example, take 3173.

Luck during qualification rounds does matter, and we found that out the hard way at Boston this weekend. Out of the 10 qualification matches we were the only team to score on our alliance 7 out of 10 times! Despite that fact we were still 6-4, but not high enough to be in the top eight, despite having one of the best robots at the regional (statistics-wise). The ranking system, however, doesn't do justice when a very good team gets poor alliance partners in qualification, and that's the one downside to a purely W-L system.

However, don't take this the wrong way and say that I want something like last years system, either. Last year was extremely wrong, as we somehow got the third seed at FLR with an absolutely terrible bot that could do nothing but score goals (but needed backup).

So the ranking system is not perfect, but I disagree that it's not too far off. This year, at least, the system should have been different.

[548swimmer]

Originally Posted by TEE

On a different note, I personally was really surprised that 548 wasn't chosen sooner... I expected them to be picked by one of the first 3 or 4 alliances.

I think this is because we didn't have a successful deployment during any of our three qualifying matches on Saturday. The first two failed, and we didn't get the chance to deploy in the third match.