Serpentine Draft

[Ian Curtis]

Originally Posted by nuggetsyl

So I am in a catch 22 I like and dislike the serpentine draft. This year IMO way to many teams have been killed by the serpentine draft. You just can't over come a broken robot.

I heard a suggestion (2590 drive coach) and I really liked it. It would add a new twist to alliance selections. Allow the 1 seed pick where they want to draft from. Then the 2 seed picks where they want to draft from and so on. It means seeding number 1 or 2 is not a as a severe penalty as it currently is allowing them to build an alliance of working robots.

Teams and the audience have a hard enough time deciphering the current drafting system. I don't think we should make it any more complex, or alliance selection along might run through the awards ceremony.

[Harman341]

Cool idea, but consider this situation

Team A 1st pick 8th seed
Team B 2nd pick 2nd seed
Team C 3rd pick 1st seed
Team D 4th pick 3rd seed
Team E 5th pick 6th seed
Team F 6th pick 4th seed
Team G 7th pick 5th seed
Team H 8th pick 7th seed

Team A can now pretty easily employ the scorched earth strategy by picking the first, second, and third seed. One of them would either have to join the alliance with the 8 seed or be left unable to pick one another. Cool idea, but I don't see why a team would allow for this possibility. Plus, this year at the 3 events I've been to, the #1 alliance made it to the finals all three times, and won twice.

[OWilliamson]

Originally Posted by Grim Tuesday

Couldn't the number one seed just decline in this case?

Yes, looks like I should think about my posts a little more before I reply.

Thanks for catching that.

[Duncan Macdonald]

This is a neat thought experiment. I expect most teams would still select the top spot available and there are often booing accusations following a decline.

Raising the minimum number of teams would also decrease the serpentine's bite but I suspect most venues are near capacity already.

# of teams | minimum elimination robot percentile

30 | 0.2
36 | 0.33
42 | 0.43
48 | 0.5

Allowing the top alliances to select their opponent from those below them each round would help the stronger alliances to advance.

[MisterJ]

Originally Posted by cmrnpizzo14

Has anyone actually done some sort of analysis this year on the number of regionals won by #1 seeds, #2 seeds, #3 seeds etc. and compared it to previous years? I am not sure that I buy that this year alone there are a standout number of upsets by low seeds.

I'd also be interested in seeing the winning percentage of each seed in each round this year. I bet this year's percentages are very much different from the past two, with lower seeds winning much more often.

[Christopher149]

Originally Posted by MisterJ

I'd also be interested in seeing the winning percentage of each seed in each round this year. I bet this year's percentages are very much different from the past two, with lower seeds winning much more often.

My immediate, poorly-scientific, non-comprehensive view is it isn't that much different than the past.

This year
#1 seed wins: Central Illinois, Palmetto, Alamo, Escanaba, Southfield, Granite State, Northern Lights, Arizona, Buckeye, Montreal, Greater DC, Hawaii
#2 seed wins: Lake Superior, Crossroads, 10k lakes
#3 seed wins: Traverse City, San Diego, Los Angeles, Livonia
#4 seed wins: Boilermaker
#5 seed wins: Greater Toronto East
#8 seed wins: Center Line
blue wins every elim series: Peachtree (this is so weird)

Devil's advocate: but isn't it a bit exciting to not always have #1 seed win?

[dodar]

Originally Posted by Christopher149

My immediate, poorly-scientific, non-comprehensive view is it isn't that much different than the past.

This year
#1 seed wins: Central Illinois, Palmetto, Alamo, Escanaba, Southfield, Granite State, Northern Lights, Arizona, Buckeye, Montreal, Greater DC, Hawaii, FLR, North Bay
#2 seed wins: Lake Superior, Crossroads, 10k lakes, Orlando
#3 seed wins: Traverse City, San Diego, Los Angeles, Livonia
#4 seed wins: Boilermaker
#5 seed wins: Greater Toronto East
#8 seed wins: Center Line
blue wins every elim series: Peachtree (this is so weird)

Devil's advocate: but isn't it a bit exciting to not always have #1 seed win?

Added a few more.

[cmrnpizzo14]

Using the most advanced pen and post-it note technology I have lazily gathered some data from week 5 this year and last year. Usual data caveats apply that this is a small sample size and whatnot.

Out of the 20 week 5 events in 2014:

1st Seed - 8
2nd Seed - 3
3rd Seed - 4
4th Seed - 2
5th Seed - 1
6th Seed - 2

Out of the 14 week 5 events of 2013:

1st Seed - 8
2nd Seed - 3
3rd Seed - 1
4th Seed - 1
5th Seed - 1

Any alliance not mentioned did not earn a victory at a regional or district event. As far as "official" statistics go this data is probably considered useless but it is fun speculation and if anyone would like to add data from another week that would be appreciated.

Notably, being the #1 seed (and pick) seems to show a clear correlation to success. The 2nd seed doesn't do bad either but past that it seems pretty even. I would need more data to really do anything.

[nuggetsyl]

1st seed Hatboro-Horsham,
2nd seed Clifton
3rd seed Mt. Olive, Bridgewater-Raritan
5th seed Springside Chestnut Hill
7th seed Lenape-Seneca

5th seed was in the finals 3 times won it once.

[MisterJ]

Quick calculation for all 2014 events, weeks 1-3

1-seed: 23
2-seed: 10
3-seed: 3
4-seed: 1
5-seed: 2
6-seed: 1
7-seed: 1
8-seed: 1

[Lil' Lavery]

There are too many issues with this system for it to be workable, and the actual changes would likely be relatively minor.

What happens when one captain accepts an invitation from another captain? Does the 9th seed automatically inherit the 8th draft position, as they do now? Wouldn't that defeat much of the purpose of changing positions? Do they inherit the vacated draft position, regardless of where they would want to select? Do all the remaining captains choose draft positions again?

More importantly, how is the bracket layout decided? Is the "#1 alliance" still the #1 seed even if the #1 seed picks in another location? What happens when the #6 captain picks the 4th ranked team, and they accept? Is it the #6 alliance or the #4 alliance? Or is the "#1 alliance" the team with the 1st selection, regardless of what rank they were? That would introduce a whole new mechanic of "gaming the system," where lower ranked captains could intentionally select draft positions to match up against other captains they feel they would fare well against. For instance, if I know the #3 captain is "weak" and they've selected to pick 5th, I'll select to pick 4th so I can play against them. This scenario would create a disadvantage for the 1st ranked team, as they would be able to chose who they're playing against.

How does this impact district ranking points for alliance selection and alliance captaincy? Does the #1 seed still receive the 16 points for being the #1 captain, or would they get the lesser points for being a lower captain? It would create further disincentive for, say, the #3 captain (14 points) to accept the, say, 5th invitation (12 points).

Few teams possess the scouting sophistication to fully take advantage of this system anyway. And given the very short time span between when rankings are finalized and alliance selection occurs, no team has enough time to truly interpret the data and alliance permutations to determine their ideal draft position.

[Chief Hedgehog]

Outcome of Minnesnowta Regionals

2014
#2 Alliance wins at 10,000 Lakes
#5 Alliance wins at North Star
#2 Alliance Wins at Lake Superior
#1 Alliance wins at Northern Lights

2013
#1 Alliance wins at 10,000 Lakes
#2 Alliance wins at North Star
#7 Alliance Wins at Lake Superior
#3 Alliance wins at Northern Lights

2012
#2 Alliance wins at 10,000 Lakes
#2 Alliance wins at North Star
#1 Alliance Wins at Lake Superior
#3 Alliance wins at Northern Lights

So a little bit of a mix up in Minnesnowta the last few years... However this may be to such a large influx of teams in the last 4-5 years in MN. Last count we are around 190(?).

[Chris Hibner]

Something similar was brought up last year: http://www.chiefdelphi.com/forums/sh...2&postcount=29, but in real-time (i.e., you don't have to declare up-front, but you can decide when you want to pick as you go depending on what the next in line decides to pick).

[Peter Matteson]

Originally Posted by MisterJ

Quick calculation for all 2014 events, weeks 1-3

1-seed: 23
2-seed: 10
3-seed: 3
4-seed: 1
5-seed: 2
6-seed: 1
7-seed: 1
8-seed: 1

1st seed wins - 54.7%
1st & 2nd seeds win - 78.6%
1st, 2nd, 3rd seed win - 85.7%

Seems about right to me.
The advantage for seeding top 3 is there as they still win a disproportionate amount.

FWIW at Hartford District there was only one upset and it was 3 over 2 in the semi which was close battle.

Have there been any "chalk bracket", aka "All Red", regionals/districts this year?

[Damiaen_Florian]

I think that as the fields of play get deeper, meaning more high level teams, the 1st seed will be much more likely to win, simply because the 3rd robot picked by the 1st seed will be closer to the skill level of the 3rd robot picked by the 8th seed. Wait until district championships and world championships and I think you'll again see the higher seeds winning as often as they normally do.