Serpentine Draft

[nickpaterni]

I've not seen any threads on this topic before, and I'm sure it has been discussed ad infinium, but I always thought that a simple change would make elimination alliance selection better from an excitement / game standpoint: The top 8 are not allowed to pick from anyone in the top 8.
This year, we've seen top alliances picking from deeper in the filed than in previous years, and I think doing that has really lent excitement to eliminations. For me, it is really exciting to see alliances with one powerhouse team face off against a similar alliance with one powerhouse team. This furthermore makes it so the 1 or 2 seed is "going" to win which is just not very exciting (unless you are in the 1 or 2 seed alliance).
Of course, there could be abuses of this system - intentionally losing qual matches to keep yourself out of the top 8, but I think a combination of GP and perhaps a tweaking of the points earned for match wins vs. draft selection points could render that problem almost null.
In the end, I think everyone can agree having an exciting elims is preferable to watching one super alliance stomping everyone else into the ground, and I think not mixing the top 8 is the way to do it.
(OK, now tell me why my idea is flawed )

[Peter Matteson]

Originally Posted by nickpaterni

I've not seen any threads on this topic before, and I'm sure it has been discussed ad infinium, but I always thought that a simple change would make elimination alliance selection better from an excitement / game standpoint: The top 8 are not allowed to pick from anyone in the top 8.
Of course, there could be abuses of this system - intentionally losing qual matches to keep yourself out of the top 8, but I think a combination of GP and perhaps a tweaking of the points earned for match wins vs. draft selection points could render that problem almost null.
In the end, I think everyone can agree having an exciting elims is preferable to watching one super alliance stomping everyone else into the ground, and I think not mixing the top 8 is the way to do it.
(OK, now tell me why my idea is flawed )

This was the one of the systems used for alliance selection in the past (2002-2003 IIRC).
It worked horribly for the reason that you mention. There was a ton of collusion because no one wanted to be stuck unable to ally with the top teams.

Even worse than this was the Auto pairing where 1/5, 2/6, 3/7 and 4/8 (2001)were automtically paired together.

[nickpaterni]

Originally Posted by Peter Matteson

This was the one of the systems used for alliance selection in the past (2002-2003 IIRC).
It worked horribly for the reason that you mention. There was a ton of collusion because no one wanted to be stuck unable to ally with the top teams.

Even worse than this was the Auto pairing where 1/5, 2/6, 3/7 and 4/8 (2001)were automtically paired together.

So do you not think that peer GP pressure would work to alleviate this problem? You would think it would be plainly obvious if collusion was happening - oh, we went 5-7-0 but then all of a sudden start scoring huge in elims - it could even be reviewed and issued a DQ / Red Card if deemed collusion

Or, what if the point values for qual matches were increased so that losing qual matches was something that in the long run needed to be avoided at all costs for you to progress past the district event?

I wasnt around in the early 2000s so I didn't know it used to be that way, very interesting.

[Jon Stratis]

Originally Posted by nickpaterni

So do you not think that peer GP pressure would work to alleviate this problem? You would think it would be plainly obvious if collusion was happening - oh, we went 5-7-0 but then all of a sudden start scoring huge in elims - it could even be reviewed and issued a DQ / Red Card if deemed collusion

Or, what if the point values for qual matches were increased so that losing qual matches was something that in the long run needed to be avoided at all costs for you to progress past the district event?

I wasnt around in the early 2000s so I didn't know it used to be that way, very interesting.

You'd be surprised at the number of teams that do better in elims than they do in quals. For example, in Logomotion my team never deployed a minibot until elims at our second regional... and that minibot made all the difference in winning the competition that year. It wasn't due to any sort of plan, just an inability to get all of the pieces working together correctly until then.

[nickpaterni]

Originally Posted by Jon Stratis

You'd be surprised at the number of teams that do better in elims than they do in quals. For example, in Logomotion my team never deployed a minibot until elims at our second regional... and that minibot made all the difference in winning the competition that year. It wasn't due to any sort of plan, just an inability to get all of the pieces working together correctly until then.

How many times during quals did you try and fail to deploy the minibot? Did you have any partial successes? My team typically also improves significantly during quals and into elims, but anyone watching can easily see that the improvement is gradual - we get a subsystem half working during a match, and in the next match it is working better than the next match it is working really well. Very different from all of a sudden being fantastic after fail fail fail.
For a team to intentionally sandbag like that would have to involve the mentors, drive team and pit crew and you would hope at least one of them would be unwilling to participate in cheating like that.

Maybe I'm being optimistic about others ability to abide by a rule that is hard to enforce and gives a potential huge advantage (bag & tag anyone?)

[Alan Anderson]

Originally Posted by nickpaterni

For a team to intentionally sandbag like that would have to involve the mentors, drive team and pit crew and you would hope at least one of them would be unwilling to participate in cheating like that.

All it takes is one or two people on the drive team to decide to play badly. A single programmer can have a very large effect on the robot's performance without anyone else's involvement. A team-wide conspiracy is not required.

[nickpaterni]

Originally Posted by Alan Anderson

All it takes is one or two people on the drive team to decide to play badly. A single programmer can have a very large effect on the robot's performance without anyone else's involvement. A team-wide conspiracy is not required.

Excellent point. Sad point, but true nonetheless. I guess we will have to rely upon GDC to create games that break that paradigm like they did this year - having a balanced alliance > 2 best scorers + box on wheels

[BBray_T1296]

Originally Posted by waialua359

Also, no offense to any of the teams we played, but IMO, I found the #8 alliance much tougher to play against in all 3 regionals we played then the semifinals matchup.

Yay! I'm a statistic!

But seriously, The #1v#8 matches have been the most intense I have seen on most webcasts, and have been the most exciting (determined by audience decibel level) of both regionals we were in the #8 alliance for. Lots of teams root for the underdog and they are not being disappointed. Even though most often the #1 seed wins, it is always a win by just a last second truss toss or the like.

[Anupam Goli]

Death by Serpentine doesn't just occur in small events. In some of the weakest events, death by serpentine is a common occurrence. In each of the last three peachtree regionals, the #1 seed has lost in quarters or semis, and twice now the #6 seeded alliance has won the finals.

Finding an inbounder at our regional this year was like trying to mine for gold. It's so hard to find someone who can hold onto the ball, release it, and play good defense. There were teams that were great at possessing the ball, but were way too top heavy and could fall over on their own. having these teams play defense would've spelled disaster. There were teams that could play defense very effectively, but couldn't manipulate the ball. Those that could do both seemed to be in the very small minority. In fact, if I erased all of the teams that could truss and were on our first pick list as well, I had a grand total of 4 teams that I would like as our 2nd pick. i had to pull some of those non-ideal partners onto our list because the depth of the event was not good at all, despite 64 teams being in attendance.

This was likely one of the reasons why some of the #1-#4 alliances lost in quarters at peachtree. (Other reasons include ref's discretion, which has been beaten to the ground already, and mechanical problems popping up). Even with a mediocre trusser and shot, if that third robot can inbound and play defense more effectively than the other alliance can, cycle times are shortened and the other alliance has to face heavier defense.

[nickpaterni]

Originally Posted by BBray_T1296

Yay! I'm a statistic!

But seriously, The #1v#8 matches have been the most intense I have seen on most webcasts, and have been the most exciting (determined by audience decibel level) of both regionals we were in the #8 alliance for. Lots of teams root for the underdog and they are not being disappointed. Even though most often the #1 seed wins, it is always a win by just a last second truss toss or the like.

Us too - we were an 8th seed alliance that did beat the 1st seed (MI-Centerline) and went on to win the event. After our first match, we had other teams coming up to us to cheer us on and our alliance's outlook was that win or lose, we felt so good about having won even one match vs 1st seed and having so many other teams rooting for the underdog. That's why balanced alliances are so good for FIRST - when everyone feels like they have a chance and the matches are close, elims are epic. Nobody really talks about the elims where #1 stomps everyone because they had the best robots and nobody else stood a chance.

[nickpaterni]

Originally Posted by Anupam Goli

Death by Serpentine doesn't just occur in small events. In some of the weakest events, death by serpentine is a common occurrence. In each of the last three peachtree regionals, the #1 seed has lost in quarters or semis, and twice now the #6 seeded alliance has won the finals.

Finding an inbounder at our regional this year was like trying to mine for gold. It's so hard to find someone who can hold onto the ball, release it, and play good defense. There were teams that were great at possessing the ball, but were way too top heavy and could fall over on their own. having these teams play defense would've spelled disaster. There were teams that could play defense very effectively, but couldn't manipulate the ball. Those that could do both seemed to be in the very small minority. In fact, if I erased all of the teams that could truss and were on our first pick list as well, I had a grand total of 4 teams that I would like as our 2nd pick. i had to pull some of those non-ideal partners onto our list because the depth of the event was not good at all, despite 64 teams being in attendance.

This was likely one of the reasons why some of the #1-#4 alliances lost in quarters at peachtree. (Other reasons include ref's discretion, which has been beaten to the ground already, and mechanical problems popping up). Even with a mediocre trusser and shot, if that third robot can inbound and play defense more effectively than the other alliance can, cycle times are shortened and the other alliance has to face heavier defense.

I would argue that a winning strategy this year is for a good shooting / scoring robot in 1-2 seed place picking the BEST inbound/truss/defense bot available, NOT the second best shooter / scorer first.

[Anupam Goli]

Originally Posted by nickpaterni

I would argue that a winning strategy this year is for a good shooting / scoring robot in 1-2 seed place picking the BEST inbound/truss/defense bot available, NOT the second best shooter / scorer first.

I didn't exactly say that these teams chose the second best shooter/scorer first. Most of the alliance captains picked partners that could play a role complementary to their own, that wasn't the issue. The issue i was trying to highlight was the lack of depth, making finding that third robot (for most alliances this was an inbounder or a robot that could hold onto the ball just for assist points who could play defense.)

[nickpaterni]

Originally Posted by Anupam Goli

I didn't exactly say that these teams chose the second best shooter/scorer first. Most of the alliance captains picked partners that could play a role complementary to their own, that wasn't the issue. The issue i was trying to highlight was the lack of depth, making finding that third robot (for most alliances this was an inbounder or a robot that could hold onto the ball just for assist points who could play defense.)

Ah, ok. Third robot still can make or break the alliance this year especially with penalties for sure.

[BBray_T1296]

Originally Posted by Anupam Goli

Death by Serpentine doesn't just occur in small events. In some of the weakest events, death by serpentine is a common occurrence. In each of the last three peachtree regionals, the #1 seed has lost in quarters or semis, and twice now the #6 seeded alliance has won the finals.

-Just because you are seeded #1 (or in the top 10) does not mean you deserve to be there. We play too few matches to get a perfectly fair schedule. Some people will get to the top (and randomly, yes, consistently) because they get good schedules. On the reverse, some good teams get bad schedules and seed poorly for no other reason. Powerhouse teams are ones who win regardless of the luck of the draw.

-Just because you are seeded #1, does not mean you have scouting worth a grain of sand. If the #6 seed knows a thing or two about each robot, then of course they are going to select more suitable partners for their alliance, serpentine or not.

-Are upsets really a bad thing? You make it sound like: "If the #1 seed is not winning, the system is broken." Would not the best solution be then to just remove the eliminations altogether? Just seed #1,2 and 3 and here's your banner? I think anyone will admit this is too far, but where is the line? The system is the way it is, it is, in actual fact, not hard to comprehend in the slightest, (My 70-something grandmother and her sister came to a regional and understood the draft immediately).

-Not one single post in this thread has actually suggested an alternative system that is
a) as "difficult" or easier to explain
b) not so one sided as to defeat the purpose of eliminations altogether
c) half as good as serpentine

[Anupam Goli]

Originally Posted by BBray_T1296

-Just because you are seeded #1 (or in the top 10) does not mean you deserve to be there. We play too few matches to get a perfectly fair schedule. Some people will get to the top (and randomly, yes, consistently) because they get good schedules. On the reverse, some good teams get bad schedules and seed poorly for no other reason. Powerhouse teams are ones who win regardless of the luck of the draw.

-Just because you are seeded #1, does not mean you have scouting worth a grain of sand. If the #6 seed knows a thing or two about each robot, then of course they are going to select more suitable partners for their alliance, serpentine or not.

-Are upsets really a bad thing? You make it sound like: "If the #1 seed is not winning, the system is broken." Would not the best solution be then to just remove the eliminations altogether? Just seed #1,2 and 3 and here's your banner? I think anyone will admit this is too far, but where is the line? The system is the way it is, it is, in actual fact, not hard to comprehend in the slightest, (My 70-something grandmother and her sister came to a regional and understood the draft immediately).

-Not one single post in this thread has actually suggested an alternative system that is
a) as "difficult" or easier to explain
b) not so one sided as to defeat the purpose of eliminations altogether
c) half as good as serpentine

- I've gone to the peachtree regional every year for the last 6 years. In the years I was referring to, the #1 seed definitely had the most dominant robot at the event. I will say this year's #6 seed at Peachtree definitely had the strongest alliance at the event, but nobody was expecting the #1 and #2 seeded alliances to lose in quarters.

-Each of these teams had great scouting. Even if they didn't, do you think that none 8 teams that were on the #1-#4 alliances this year didn't have the scouting to make a good 2nd pick? I know that great scouting is essential, because scouting cost us 2 regionals last year. Trust me, each of the teams tried to find the best partner, but after 1683 and 832 were gone, the list of good 3rd robots was very small.

-Upsets aren't a bad thing, the teams that build the greatest alliances SHOULD be rewarded. The problem is, like I said, at weaker regionals, building the best alliance is often only possible in the lower seeds. This encourages sandbagging a match to get a lower seed, especially in a game like aerial assist, where like others have said, the top 2 scoring robots aren't exactly the best alliance.

I personally don't think changing the way alliance selection happens can help mitigate this issue, I just wanted to bring up that Death By Serpentine isn't limited to small events, and in fact exists at larger events. The end solution would be to bring up the level of competition, but that's another discussion.