Chief Delphi

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)
-   Rules/Strategy (http://www.chiefdelphi.com/forums/forumdisplay.php?f=6)
-   -   Working with alliance partners (http://www.chiefdelphi.com/forums/showthread.php?t=149102)

ns3517 21-06-2016 22:54
Working with alliance partners
 
I'm curious how different teams strategise and how they work with alliance partners in qualification and playoff rounds.

frcguy 21-06-2016 23:00
Working with alliance partners
 
We meet with both teams about half an hour before qual matches to discuss strategy and who we're playing. Then we usually go over everything in queue to make sure we're on the same page.

evanperryg 21-06-2016 23:38
Re: Working with alliance partners
 
Quote:
Originally Posted by frcguy (Post 1593800)
We meet with both teams about half an hour before qual matches to discuss strategy and who we're playing. Then we usually go over everything in queue to make sure we're on the same page.
Same for us, mostly. At champs, we also started going around about 45-60 minutes before queue, asking partners if there was anything we could do to help them be ready for the match. This proved to be extremely helpful, and helped keep our ranking from being too dismal.

Jay O'Donnell 22-06-2016 00:12
Re: Working with alliance partners
 
The best thing I've found is to ask each of your partners what they want to do and try to work around them, as long as it makes sense strategically. Teams generally do better when they're doing something they want to do. Above all else, be kind to your partners and they'll probably work well with you.

smitikshah 22-06-2016 00:48
Re: Working with alliance partners
 
Pretty much the same as the other responses. We go over about 30 minutes before the match and compile our scouting information to efficiently strategize against the opposite alliance as well.

bkahl 22-06-2016 01:38
Re: Working with alliance partners
 
I would be interested in how other teams deal with potentially over-zealous (for lack of a better word?) qualification partners. By that I mean teams who say they can do something you know they really can't.

For example- in 2014, Taking a ball away from another team in autonomous to run a 2(+) ball auto

How do others remain GP and respectful to other teams while convincing them NOT to do something they really want to do.

In the past I have tried the following:

1) Presenting them with Raw Data to prove their idea is not necessarily the best. I have found this to be very blunt, and while it usually gets the job done, sometimes telling a team they haven't yet scored in auto isn't the most fun thing to do.

2) Playing nice and trying to slowly convince them that while their idea is "good", there may be a better alternative. This usually has less success as pretty much everyone likes to stick to their guns.

frcguy 22-06-2016 01:38
Quote:
Originally Posted by evanperryg (Post 1593804)
Same for us, mostly. At champs, we also started going around about 45-60 minutes before queue, asking partners if there was anything we could do to help them be ready for the match. This proved to be extremely helpful, and helped keep our ranking from being too dismal.

Yeah, we did that too. As an example, we learned before qual 79 at SVR that our alliance partner 253 was getting disabled because of their battery exiting their robot. We sent over a few students, a mentor, and a cart full of tools, and by the time the match rolled around we set them up with a new plate to hold in their battery. It was great being able to help out another team, and they were able to breach defenses without getting disabled.

pabeekm 22-06-2016 03:59
Re: Working with alliance partners
 
Quote:
Originally Posted by bkahl (Post 1593814)
I would be interested in how other teams deal with potentially over-zealous (for lack of a better word?) qualification partners. By that I mean teams who say they can do something you know they really can't.
One good way to work with overconfident teams is to negotiate strategies that change over the course of the match. For example, in 2014 lots of teams overestimated their ability to shoot and wanted to use the high goal. Some easy ways to compromise with them were 1) ask that the alliance complete a few cycles using the low goal at the start of the match or 2) negotiate what would happen if shooting in the high goal failed a certain number of times.

Those kinds of compromises don’t apply to one-time things like auto, but they can resolve tele-op disagreements amicably.

Basel A 22-06-2016 08:07
Re: Working with alliance partners
 
Jay's post above hits the most important point. The first question to ask is "what would you like to do?" Then discuss between all three teams to make the three answers work together.

Quote:
Originally Posted by bkahl (Post 1593814)
I would be interested in how other teams deal with potentially over-zealous (for lack of a better word?) qualification partners. By that I mean teams who say they can do something you know they really can't.

For example- in 2014, Taking a ball away from another team in autonomous to run a 2(+) ball auto

How do others remain GP and respectful to other teams while convincing them NOT to do something they really want to do.

In the past I have tried the following:

1) Presenting them with Raw Data to prove their idea is not necessarily the best. I have found this to be very blunt, and while it usually gets the job done, sometimes telling a team they haven't yet scored in auto isn't the most fun thing to do.

2) Playing nice and trying to slowly convince them that while their idea is "good", there may be a better alternative. This usually has less success as pretty much everyone likes to stick to their guns.
Great question, but a tough one to answer because it's always a different situation. But there's usually a strategy for mitigating their, uh, over-zealousness. For shooting in 2014, shot accuracy was a pretty strong argument, but tone conveys as much as the actual data. For arguing over who shoots, get to the point where you both equally want to shoot, then check the scouting data. You say "Hmm, looks like our shot percentage is a bit higher..." and let them come to the conclusion themselves. For arguing over auton balls, I'd invoke fairness. Nobody gets an extra ball unless another team wants to give it up.

Sometimes you can just leave them be. I recall a situation in 2015 when one of our partners said they could make 2-3 stacks of totes, and the other partner said they could cap 2-3 stacks of totes. We knew for sure neither team was representing their output accurately, but suggested they work together to make capped stacks because regardless, that's the optimal strategy. I think they ended up with one capped stack, and nobody's feelings got hurt.

The comment from Patricia above is great because it uses their own over-confidence. When you say, "if you miss, let's go low goal for the rest of the match," they're thinking "no way we're gonna miss." Nobody's feelings get hurt and you're really not far from optimal strategy.

Probably over 90% of the time, the match plan negotiated by the strategists on our team ends up being the strategy for the match. In cases where you're trying to slowly move a partner from their plan to yours, "the drive coach disagrees with the strategists" is a great excuse to change things up.

Eric Scheuing 22-06-2016 08:49
Re: Working with alliance partners
 
Quote:
Originally Posted by bkahl (Post 1593814)
How do others remain GP and respectful to other teams while convincing them NOT to do something they really want to do.
If you are the alliance captain, you should get the final say in the matter. You built your alliance to run a strategy you already had planned (hopefully), and you may need to remind a team that you picked them for that reason. Typically I would say/like to be told "We picked you because you do ______ very well, and it is an important part of the strategy we want to use."

However, if a team has a good strategy figured out for the whole alliance rather than just something new they want to try on their own, it's probably worth considering. Again, usually teams get picked because what they do well compliments a pre-determined strategy, and asking the whole alliance to do something new, even if it sounds like a killer strat, may result in a team doing something they have little to no experience in and could fail spectacularly... or not.

EDIT: As for quals, I feel each team should be given the opportunity to play however they want to for that match. IMO, winning is secondary to playing well in quals. If a team member wants to employ a new strategy that they can execute on their own, then I have no problem with letting them try it out. However, if it requires us to do something we aren't comfortable with or well practiced at, I have no problems with saying "sorry, we're not going to do that." Our coach has a mantra: "play your game." We know the game we are going to play, and it's more important to us to do what we do well consistently than to try something new every match.

Hitchhiker 42 22-06-2016 09:00
Re: Working with alliance partners
 
Quote:
Originally Posted by bkahl (Post 1593814)
In the past I have tried the following:

1) Presenting them with Raw Data to prove their idea is not necessarily the best. I have found this to be very blunt, and while it usually gets the job done, sometimes telling a team they haven't yet scored in auto isn't the most fun thing to do.

2) Playing nice and trying to slowly convince them that while their idea is "good", there may be a better alternative. This usually has less success as pretty much everyone likes to stick to their guns.
The way our team approaches it is to suggest another alternative that, in the other team's mind, is just as good, but also has them not doing the thing they can't do.

In general, we try to go over as many possibilites of a good gameplay as possible, and choose the best scenario, which we try to stick to.

Chris is me 22-06-2016 09:51
Re: Working with alliance partners
 
Quote:
Originally Posted by bkahl (Post 1593814)
I would be interested in how other teams deal with potentially over-zealous (for lack of a better word?) qualification partners. By that I mean teams who say they can do something you know they really can't.

For example- in 2014, Taking a ball away from another team in autonomous to run a 2(+) ball auto

How do others remain GP and respectful to other teams while convincing them NOT to do something they really want to do.
This is among the hardest things to do in FRC strategy, and to some extent the answer varies depending on the team and region. In one region, you might be a reasonable and level headed coach, when in another region, you might be considered an arrogant jerk. It's context dependent as well.

Quote:
1) Presenting them with Raw Data to prove their idea is not necessarily the best. I have found this to be very blunt, and while it usually gets the job done, sometimes telling a team they haven't yet scored in auto isn't the most fun thing to do.
This is kind of the nuclear option, and something that you should use mainly when a partner is lying about their capabilities. Opening with this, when you don't know yet for sure if a team is going to be stubborn / arrogant, can give off the wrong impression.

----

So here's my general approach to this.

Open the discussion by asking each partner to say what their robot is capable of and what they like to do in a match (including yourself). Then discuss the needs of the alliance for the match, what does the alliance in a general sense (not specific robots) need to do to win the match / achieve your objectives - "we need to not miss any balls in auton" "we need to get the breach point and I think we can get the capture". Then you kind of suggest a strategy that uses bits and pieces of what everyone said about their own capabilities ("since you guys are also good at low goals, what if you focused on that while we took the high goal?").

Sometimes you'll encounter pushback, at which point you don't have much choice but to find compromise, at least early in the day. Try to find a way that they can do their thing while also doing the thing you want. "What if you play defense for the first minute, then come back early and score a few goals at the end?" That sort of thing. If they are flat out saying they are capable of things they aren't, that's when you bust out the scouting data, and you do it in a concerned way rather than an accusatory way. If they are confident THIS is the match they will hit their auto, ask them what they have changed since the last match. Ultimately, you can't make a partner do something they don't want to, and if you push too hard they will just not do the task you asked of them anyway.

Some teams are just impossible to work with, and will only do what they want to do no matter what you ask of them. All you can really do is suck it up and put them on your DNP list for alliance selections and move on. Try not to be one of those teams, even if you're at the top tier.

notmattlythgoe 22-06-2016 10:02
Re: Working with alliance partners
 
Quote:
Originally Posted by Chris is me (Post 1593833)
This is among the hardest things to do in FRC strategy, and to some extent the answer varies depending on the team and region. In one region, you might be a reasonable and level headed coach, when in another region, you might be considered an arrogant jerk. It's context dependent as well.



This is kind of the nuclear option, and something that you should use mainly when a partner is lying about their capabilities. Opening with this, when you don't know yet for sure if a team is going to be stubborn / arrogant, can give off the wrong impression.

----

So here's my general approach to this.

Open the discussion by asking each partner to say what their robot is capable of and what they like to do in a match (including yourself). Then discuss the needs of the alliance for the match, what does the alliance in a general sense (not specific robots) need to do to win the match / achieve your objectives - "we need to not miss any balls in auton" "we need to get the breach point and I think we can get the capture". Then you kind of suggest a strategy that uses bits and pieces of what everyone said about their own capabilities ("since you guys are also good at low goals, what if you focused on that while we took the high goal?").

Sometimes you'll encounter pushback, at which point you don't have much choice but to find compromise, at least early in the day. Try to find a way that they can do their thing while also doing the thing you want. "What if you play defense for the first minute, then come back early and score a few goals at the end?" That sort of thing. If they are flat out saying they are capable of things they aren't, that's when you bust out the scouting data, and you do it in a concerned way rather than an accusatory way. If they are confident THIS is the match they will hit their auto, ask them what they have changed since the last match. Ultimately, you can't make a partner do something they don't want to, and if you push too hard they will just not do the task you asked of them anyway.

Some teams are just impossible to work with, and will only do what they want to do no matter what you ask of them. All you can really do is suck it up and put them on your DNP list for alliance selections and move on. Try not to be one of those teams, even if you're at the top tier.
What makes this even harder is when you/a partner are trying to showcase capabilities to be picked towards the end of the event. Especially when it does not necessarily fall into the best strategy for the alliance.

Michael Hill 22-06-2016 10:09
Re: Working with alliance partners
 
This year, we ran strategy a bit different. We spent a lot of time thinking about each match our team was in, delivered a strategy suggestion (including an honest overview of all the teams in the match, defense suggestions, etc.) to the coach and they generally kept to the suggestion. When speaking to the other teams, our drive team was armed with real-time scouting data and the strategy suggestion. Sometimes its easiest to not let people get overzealous in the first place. When approaching, our drive team will say something like, "hey, let's talk about the next match, we see you guys can cross [X] defense pretty well in autonomous, would you like to do that while the other team does [Y] and we do [Z]?" Much of the time, I found all the drive teams are happier when presented with raw scouting data. Sometimes they get overzealous because they don't really know how they actually perform.

Also, being 100% transparent with your own abilities is also key to developing a good relationship. If you know you have a difficult time crossing the ramparts, say that you draw vacuum at it. Complements go a long way when talking to other drive teams. Don't just say "hey, you draw vacuum, go play defense". With scouting data, you can say, "hey, could you guys do [the one defense they're good at] while we go score boulders? Team 12345 is really good at scoring boulders, but they're easily defended and need to be stopped to win the match. Just go push them up on the batter, then when they come after you again, push them back up."

When you put it in the context that the team playing defense is actually contributing to the match win, the "go play defense" is a much easier pill to swallow.

Eric Scheuing 22-06-2016 10:11
Re: Working with alliance partners
 
Quote:
Originally Posted by notmattlythgoe (Post 1593837)
What makes this even harder is when you/a partner are trying to showcase capabilities to be picked towards the end of the event. Especially when it does not necessarily fall into the best strategy for the alliance.
When it gets to the point where you're showcasing abilities to be picked, winning each match probably doesn't matter as much as actually showing off what you need to show off.

Of course, if one of your alliance partners needs the win to be seeded or something, you should try to help them out as much as possible.

ns3517 22-06-2016 11:23
Re: Working with alliance partners
 
Thanks for all the replies guys! If anyone has more comments feel free to post.

Jcarbon 22-06-2016 15:09
Re: Working with alliance partners
 
I agree with everything that has been said so far. Two specific things that make strategy planning easier and more efficient:

1. Scouting Data - Before planning each match, the scouting team prints out a data sheet for the strategists. The sheet has relevant info on all 6 teams, color coded and organized for easy reading.

2. Whiteboard - We laser engrave a small whiteboard with the field layout to help plan the match. It's a simple addition, but it helps everyone visualize the plan.

frcguy 22-06-2016 15:49
Quote:
Originally Posted by Jcarbon (Post 1593871)
2. Whiteboard - We laser engrave a small whiteboard with the field layout to help plan the match. It's a simple addition, but it helps everyone visualize the plan.

Would you mind sharing the file you used to laser engrave the whiteboard? Ever since champs I've wanted to laser engrave my whiteboard but I couldn't figure out how to get a drawing of the field.

smitikshah 22-06-2016 15:59
Re: Working with alliance partners
 
Quote:
Originally Posted by Jcarbon (Post 1593871)
2. Whiteboard - We laser engrave a small whiteboard with the field layout to help plan the match. It's a simple addition, but it helps everyone visualize the plan.
We do the whiteboard too and it really helps. We didn't laser engrave it though we just got the field printed as a poster at Staples, and then sandwiched it between two pieces of plexiglas and just closed it by bolting the four corners. Works like a charm and only like $15.

Ben Martin 22-06-2016 16:01
Re: Working with alliance partners
 
For Quals, you have to be OK with whatever impressions of your team are given during the negotiation, as the drive team is the most highly visible window into the attitude of other teams. We DNP teams that are prohibitively difficult for us to work with, and we expect others will do the same if they feel they can't work with us (so we try to mitigate that impression as much as possible, while still trying to optimize the match for both our performance and the rankings).

Bad blood develops when two teams want to do the same thing and can't, and fortunately this year was easier than most for multiple teams to score at the same time. Pre-match negotiations for Quals went very smoothly as a whole this year (at least for us).

It's in your best interest to have your partners do the thing they will be best at if it makes sense with the match flow, since that will likely result in a higher score and more RP overall.

For elims, we're usually a lot more aggressive about pushing the strategies we believe are correct. Everyone on your alliance should have the same goal at that point, and if they don't, you probably want to avoid picking them in the first place.

GKrotkov 22-06-2016 16:29
Re: Working with alliance partners
 
I have some other thoughts on this topic, but they're much too nebulous right now to really put into a post.

However, here's a thread from 2014 that relates to a topic that was touched on here: the tradeoff between "best" strategy vs. everyone doing what they want/are comfortable with.

https://www.chiefdelphi.com/forums/s...d.php?t=127653

Of course, the best strategists in FIRST are those that manage to both have teams playing their game as well as integrating that into a cogent larger strategy - it's not necessarily a tradeoff.

Zebra_Fact_Man 22-06-2016 16:56
Re: Working with alliance partners
 
Came in to post something, read the thread, everyone already posted what I was going to say multiple times over.

NShep98 24-06-2016 23:43
Re: Working with alliance partners
 
One thing I have to add is when you have overlapping strategies, i.e. 2+ robots good at the same thing. In some cases, you just have to cut your losses because there are few ways around it. In other cases, someone may have to fall back on a weaker secondary strategy because one robot is simply better at the main one. The latter was a lot more common in Stronghold where there were more things to do, and if you find yourself in the position that you are the strongest at whatever strategy is being conflicted over, don't be afraid to be insistent on being the one to do it, within reason of course.

For example, our Stronghold bot did breaching, then defense if there was time, and had no shooting mechanism. Therefore, if we got our ideal alliance of two shooters, we would urge them to cycle through Low Bar and/or a static defense and leave the rest entirely up to us, whereas they otherwise might try to take a few other defenses, using up more time.


All times are GMT -5. The time now is 07:10.

Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi