NY Tech Valley Robot Rumble - October 4

[Chris is me]

Originally Posted by Kevin Leonard

I know for a fact 20 has a practice bot, so might 2791, 3044, and 1493. I don't know whether any of these teams have the ability to field a second drive team when many of them are likely to be helping to run the event with volunteers, and I don't know what condition these robots are in.

Our practice robot is mostly in pieces now, so we won't be bringing it and a second drive team. We had to further gut it this week to keep the comp bot alive another week. Hopefully next year.

My apologies if I was derailing the thread too much with discussion of the mechanics of 2v2; I guess I'm just excited.

[1493kd]

1493's practice bot is all apart as well. I think the 2v2 matches in Elims should be seriously considered. And then going to 3 team alliances for eliminations.

Possibly 4, 3 team alliances. Seeds 1-4 cant pick each other.

Or

5- 3 team alliances, Seed 1 can accept a bye thru to the 2nd round or get the 1st pick. If they take the bye they then get 4th pick....same deal for 2nd seed etc. 4th seed has to take the bye if it gets to them.

Just brainstorming, slow day at work..

[Kevin Leonard]

My comment about practice bots didn't only apply to the teams I listed. I know other teams coming have practice bots as well. I think it would be nice to have as many robots as possible at this event.

[PayneTrain]

Originally Posted by Mark McLeod

That might work. The dummy teams will be ranked, but teams could just overlook them.

You could DQ the dummy teams in every match and they'll show up at the bottom, at least.

[BobbyVanNess]

As far as I am aware, one of the goals for the event is to get every team 10 matches. If we did 2v2 in quals, it would mean making the event substantially longer. Also, then changing the game entirely for eliminations would throw all of these rookie drivers and refs/scorekeepers off for when it matters. Also consider that backups will likely be needed for elims. If every team is on an alliance then there may not be anyone on reserve.

Maybe we could instead be a little relaxed on turn around times between matches or have a driving frame as a qualifications reserve. If I'm correct, this is what battlecry at WPI has done. I just personally disagree with running 2v2 matches, and especially if it's not going to be consistent throughout the event. I think it just takes away from the "officialness" of the competition by straying from the actual game.

[MaxMax161]

I was with 5254 when they played at the HERE (Hoboken Engineered Robotics Event) offseason competition where we played 2v2 matches through all of quals and then did 3v3 in eliminations. It was a ton of fun, and while different from typical competition I think the 2v2 matches were actually more fun to play in. The strategic decisions are a bit different but overall simpler which I think is great for teams training/testing new drivers. I personally enjoyed the change of pace from the typical 3v3 game. It was a 15 team competition and even with 2v2 matches we could barely keep up with the pace. It was a bit hectic but a lot of fun, with 3v3 matches in this event I think it would run teams way to hard.

Originally Posted by bobby5150

I think it just takes away from the "officialness" of the competition by straying from the actual game.

This is an offseason competition where most teams will be running new drivers and pit crew with robots nearing the end of their life cycle, it wasn't official to being with.

2v2 matches are a way to get a lot more playing robots and having fun in and having played in them before I highly recommend them for this size and kind of event.

[PayneTrain]

As someone who has competed in a 2v2 offseason, 2v2 in qualifications can be great. You're not doing an offseason to have another official competition. However, you are commonly running new drive teams and pit crews. With a slowed match pace with 2v2 and fewer machines on the field, you're not throwing any new people into the inferno of a regional match and match cycle.

However, be intelligent when it comes to structuring eliminations. I don't want to detail to you how the old Robot Rumble in Richmond handled 2v2 elims for fear someone will adopt that idea in the future and credit/blame me for it.

[Chris is me]

First I do want to clarify; I'm really excited about this event and appreciative of all of the hard work people have put into planning it. If we play 3v3 matches I don't think it will ruin the experience for us or anything like that. My intent here is just to look at some data to compare the two options; I really don't want to step on any toes or otherwise come across as making demands that would alienate anyone. Can't wait for the Rumble!

All of that said, I think there's a lot of merit to playing 2v2 than 3v3, at the cost of ~2 matches per team. Whether or not this is worth it is definitely up for discussion. For the purposes of this post, I'm assuming 16 "teams" (14 real robots + 2 practice bots). All schedules I generated for this were done in MatchMaker assuming 16 teams, aiming to avoid back to back matches, and the best "quality" possible as defined by the software.

First things first: The number of matches. 10 matches per team is a noble goal and I completely support giving teams as many matches as physically possible. However, 10 matches is completely impossible with the current schedule. To give each team 10 matches, let's assume perfectly ideal conditions. Currently we are playing matches from 9:45 - 12 and from 12:30 to 1:30. This allows us 195 minutes to run these matches. 10 matches per team means 27 matches. Given each match is 2:30 long, this gives us only a 4:45 turnaround between matches! This is completely unrealistic even for a regular event, but for a 16 team event with teams going back to back, 4:45 is just not going to happen. It simply will not work. We either need more time or we need fewer matches.

So how many matches can we possibly do? Given a 7:30 turnaround time and a 2:30 duration for each match, then that's 10 minutes a match. In 195 minutes, that's ~19.5 matches. For 3v3 with 16 teams, you can give everyone 7 matches in 19 rounds. For 2v2 with 16 teams, you can give everyone 5 matches in 20 rounds. For any amount of time near 200 minutes, you're going to get two more matches with 3v3 versus 2v2. This is a significant tradeoff.

Now that we have a ballpark number of matches, I generated two schedules using MatchMaker to compare their quality. They're attached if you're curious. In the 3v3 schedule, the minimum delta between matches is 1 (back to back) for 12/16 teams; a minimum delta of 2 (1 match in between) exists for 4 teams. For the most part teams play with and against about 75% of the event in quals. For 14/16 teams, they will play with / against at least one team 4 times.

Looking at the 2v2 schedule, no teams play any back to back matches. The minimum delta between matches is 2; 11/16 teams deal with a one match turnaround. Every team partners with 5 other teams and plays against 10 other teams. No team repeats playing with the same team or against the same team, though almost every team will interact with a team twice, once on each side.

The 2v2 schedule is a lot nicer, but 2 fewer matches is a lot.
---

As for the differences between quals and eliminations, I really don't think it's a big deal. It's about as big of a change as the coopertition bridge vs triple balancing in 2012. Trusses, assists, and goals are the same; all that changes is who's on the field. It's really not "completely different strategy" to skip the inbound / pass-back robot. That said, 3v3 would still be a lot of fun and allow for a bit more... "variability"... in alliance partners.

[Steve Kaneb]

Originally Posted by Chris is me

First things first: The number of matches. 10 matches per team is a noble goal and I completely support giving teams as many matches as physically possible. However, 10 matches is completely impossible with the current schedule. To give each team 10 matches, let's assume perfectly ideal conditions. Currently we are playing matches from 9:45 - 12 and from 12:30 to 1:30. This allows us 195 minutes to run these matches. 10 matches per team means 27 matches. Given each match is 2:30 long, this gives us only a 4:45 turnaround between matches! This is completely unrealistic even for a regular event, but for a 16 team event with teams going back to back, 4:45 is just not going to happen. It simply will not work. We either need more time or we need fewer matches.

Chris brings up some good points. 2v2 or 3v3 is a pretty significant change to the schedule, but the changes to the game possibly don't matter that much to an offseason (I think driving time in matches is fundamentally the same regardless of strategic implications).

Claiming that 4:45 is an unrealistic turnaround time is false, and not the way that match turnaround times are usually described. Most events in season run sub-7 minute turnarounds (from start of match to start of match). It is not especially ambitious to run turnarounds that are a bit over 7 minutes. Don't forget that, aside from some practice robots, these robots have mostly run enough matches to make connection to the field old hat.

[Chris is me]

Originally Posted by Steve Kaneb

Chris brings up some good points. 2v2 or 3v3 is a pretty significant change to the schedule, but the changes to the game possibly don't matter that much to an offseason (I think driving time in matches is fundamentally the same regardless of strategic implications).

Claiming that 4:45 is an unrealistic turnaround time is false, and not the way that match turnaround times are usually described. Most events in season run sub-7 minute turnarounds (from start of match to start of match). It is not especially ambitious to run turnarounds that are a bit over 7 minutes. Don't forget that, aside from some practice robots, these robots have mostly run enough matches to make connection to the field old hat.

Whoops, I feel really, really dumb now! I'll edit all of that out soon.

In that case, plenty of time for 10 matches with 3v3. If we switched to 2v2, we'd have to play 7 matches. A little bit different... My points about how match scheduling works I think are still valid, but 10 matches in a 16 team event isn't ambitious for 200 minutes. My mistake.

[ThunderousPrime]

I don’t advocate a 2v2 qualification route for various reasons. Firstly, the scoring of aerial assist was designed to play as a 3v3 game both on the basis of teamwork as an attribute as well as the strategic implications of the 30pt-3 assist bonus. Assuming the scoring the values remain unchanged (see two paragraphs below) the maximum amount of points that can be scored per cycle is 30pts. I think most 2v2 cycles though will be we worth 20 or 21pts though because it is unlikely that all teams will be able score high and truss efficiently. I find the amount of points scored per cycle to be significant because teams that miss 1 or 2 auto balls compared an alliance that makes 2 or 3 auto balls will be 1 or more cycles behind.

Ex. Red Alliance misses 2 auto balls, but gets 10 pts for the mobility bonus. Blue alliance scores 2 auto balls both not hot for 30pts and gets 10 mobility points as well. Advantage Blue 40-10. At this point they are one 30pt cycle behind and would presumably clear the auto balls for one point. In 3v3 a 30pt deficit is much more manageable because 3 assists will immediately catch the Red alliance up in this scenario as well as providing one of the three robots the opportunity to defend the Blue alliance’s cycle. This scenario assumes that both scored balls for blue are “not hot” too; if either of the 2 balls scored is hot the red alliance is behind by more than a complete cycle. Also if 3 balls are scored in auto then the alliance will be almost 2 cycles behind. In short I think the 2v2 format makes comebacks very difficult if one alliance flops during autonomous and the other puts 2-3balls in the high goal.

Additionally, changing the scoring values will make a lot of more work for scorekeepers who would likely have to add assist points to the score generated by the FMS. This is the event’s inaugural year and I think that keeping with the standard FMS team randomization for quals and assist scoring will help the event flow more easily. As Bobby already mentioned the change from 2v2 to 3v3 from quals to elims will inevitably cause confusion with drivers and scorekeepers.

Another aspect of why I support stay with the standard 3v3 matches is that not all teams registered are not watching/posting on this forum. Half of the the teams registered did not post in this forum so they may or may not know that 2v2 matches are considered. I think that most teams registered with the intent to play 3v3 because it simulates a real regional or district in terms of the number of teams on the field.

In regards to 2v2 being a better format to train new drivers, I respectfully disagree.While driving on the field may be “easier”, because of fewer robots on the field and defense, I think that exposing new drivers drivers to defense and a more cluttered field presents a more realistic challenge that they would see at regionals and districts. (Realistically, no shows rarely happen at the regionals have attended. My mentality to approach offseason events is that they should be a stepping stone to get both new and veteran drivers experience for the upcoming competition season and not as a way to learn driving the robot for the first time.) I also think that the 2-3 more matches that a 3v3 format will provide will allow for more opportunities for new drivers to work out the problems that come with a very brief 2min 30sec match.

Sorry for writing a book.

[Kevin Leonard]

Originally Posted by ThunderousPrime

Ex. Red Alliance misses 2 auto balls, but gets 10 pts for the mobility bonus. Blue alliance scores 2 auto balls both not hot for 30pts and gets 10 mobility points as well. Advantage Blue 40-10. At this point they are one 30pt cycle behind and would presumably clear the auto balls for one point. In 3v3 a 30pt deficit is much more manageable because 3 assists will immediately catch the Red alliance up in this scenario as well as providing one of the three robots the opportunity to defend the Blue alliance’s cycle. This scenario assumes that both scored balls for blue are “not hot” too; if either of the 2 balls scored is hot the red alliance is behind by more than a complete cycle. Also if 3 balls are scored in auto then the alliance will be almost 2 cycles behind. In short I think the 2v2 format makes comebacks very difficult if one alliance flops during autonomous and the other puts 2-3balls in the high goal.

Then don't miss auto

Originally Posted by ThunderousPrime

Another aspect of why I support stay with the standard 3v3 matches is that not all teams registered are not watching/posting on this forum. Half of the the teams registered did not post in this forum so they may or may not know that 2v2 matches are considered. I think that most teams registered with the intent to play 3v3 because it simulates a real regional or district in terms of the number of teams on the field.

I also think the teams that registered didn't necessarily want back to back matches. It's not nice, I guarantee you that.

Originally Posted by ThunderousPrime

Sorry for writing a book.

It's CD, we're used to it.

I can see both sides of the coin, however with the small pool of teams we have, a 2v2 Qual schedule would allow for more time between matches, as well as an interesting perspective on Aerial Assist I'd like to see play out.

I think without rule changes, 2v2 Aerial Assist would make the best shooting teams more likely to sit atop the rankings at the end of the day, making for more interesting eliminations. That's my hypothesis, but I could be wrong, and there are arguments that having the best robots atop the rankings isn't necessarily the idea.

I'll enjoy the event either way. I can't wait to see you all Saturday!

[ThunderousPrime]

Originally Posted by Kevin Leonard

Then don't miss auto

True but you cannot control your alliance partner's auto.

Originally Posted by Kevin Leonard

I also think the teams that registered didn't necessarily want back to back matches. It's not nice, I guarantee you that.

True the best remedy in my opinion for that problem is to have longer than usual turnaround times.

Originally Posted by Kevin Leonard

I can see both sides of the coin, however with the small pool of teams we have, a 2v2 Qual schedule would allow for more time between matches, as well as an interesting perspective on Aerial Assist I'd like to see play out.

I don't really understand why why people find the 2v2 game play intriguing. Is it because of this specific game or do you just like the idea of a two team alliance?

Originally Posted by Kevin Leonard

I think without rule changes, 2v2 Aerial Assist would make the best shooting teams more likely to sit atop the rankings at the end of the day, making for more interesting eliminations. That's my hypothesis, but I could be wrong, and there are arguments that having the best robots atop the rankings isn't necessarily the idea.

I think the best shooting teams would definitely be ranked the highest in 2v2 rule changes or without. This is because the basic match strategy for 2v2 is inbound, truss, robot 2 acquires the ball, robot 2 scores high. Lather, rinse and repeat. There will be basically no strategic variation in 2v2 in my opinion which is why it seems very bland to me.

Additionally if you want to see how a 2v2 matches play out I recommend you watch week 1 regionals from this year. Typically, 1 robot per alliance did not touch the ball at all during the match which is similar in my mind to a 2v2. Here is the match footage from Inland Empire (thanks 1678!) if you are interested.

Originally Posted by Kevin Leonard

I'll enjoy the event either way. I can't wait to see you all Saturday!

I can't wait for the event either but I would prefer to see all matches to be 3v3! And then Ruckus 2 weeks after!

[Chris is me]

My apologies for sparking such a contentious thread! I didn't mean to cause so much controversy, I was just proposing an idea. 3v3 will be fine.

I think the point raised about not all teams being on CD, and every team having signed up for this as a 3v3 event, are more valid than any arguments of improvements over 2v2.

All of that said, I've got responses to pretty much everything that's been said lately, but in the interests of keeping the thread on track I'll keep this brief. PM me if you'd like to debate this privately. In short, 3v3 Aerial Assist didn't have a lot of strategic depth anyway (three assist cycles, do you truss or do you not, where does your inbounder sit), and with 20 points for 2 assists instead of 10 points the new game has about the same amount of depth (two assist cycles, do you truss or do you not, do you truss to the HP and IB to the low goal or do you inbound downfield and have your shooter hit the high goal and/or truss). The bigger auto deficit could be solved just by taking a ball off the field in auto, then it's only possible to score 2/3rds the points.

I based my (earlier) opinion off the experiences of a teammate playing at a 2v2 event, competing at Finger Lakes in the elims with a largely 2-assist cycle strategy, and my general desire to give my students more time to strategize and repair an old robot. If we get more than 5 minutes between matches we can make it work but as of now I am absolutely sure we're going to be rushed constantly. I guess that's part of the fun.

[MooreteP]

I'm hoping for 2 v 2. Old skool, a la 2004!

The game will still play out and Game Announcing will be easier

Kevin Leonard said it best.

Longer turnaround times = more "Cotton-eyed Joe"

I am very much looking forward to a beautiful day in Ballston Spa.
http://en.wikipedia.org/wiki/Ballston_Spa,_New_York
Make it rain!