FRC Blogged - What Do You Think? The ‘Invite to Decline’ Strategy

[Mr. Lim]

A lot of people still aren't familiar with the most important aspect of the "Scorched Earth" strategy:

The fact that you save your top pick till the end, AFTER you've scorched everyone else...

To maximize the strategy, you need to "scorch" everyone else's alliance, but still build the best one you possibly can. You need to communicate with your top pick, and let them know you intend on breaking up everyone else, and hopefully you convince them that after having done so, you can now form an alliance capable of winning the event.

To illustrate, lets say you are a "weak" #1 ranked team.

Teams ranked #2-#9 are all very strong and ranked in order of strength. You also know all these teams would rather form their own alliances, than accept your invitation.

You need to pick #3-#9 and have them all decline, THEN pick #2 (your top pick) last.

That gets you the best team at the event on your alliance, and 3-9 are all unable to pair up.

In a lot of cases, this break-up is enough to make a #1 and #2 alliance the best possible at the event. An alliance of #1 and #2 might be better than the alternative #2 and #11 - and if so, it would probably be better than all the other possible alliances from #3-#9.

"Scorched Earth" is NOT just about screwing everyone else, it's more about giving yourself the best chance at winning by selectively picking, and taking full advantage of the #1 ranking you've earned.

What amazes me is that this doesn't happen more often - that there is a negative stigma attached to it - and that a lot of people are still VERY resistant to it - likely because they haven't had the ENTIRE process explained to them!

It's takes some communication, co-operation, a bit of trust between teams, and the ability to see a few steps beyond what's immediately in front of you.

I wouldn't take away the opportunity for this strategy to happen, because I personally think it's a pretty great and ingenious way to build some pretty important critical thinking, reasoning and relationship building skills.

Despite the poor optics of a decline, and the requisite booing that inevitably happens, I think the opportunity for co-opertition and learning that goes on behind the scenes is far more valuable.

[Nick Lawrence]

I wouldn't take away the opportunity for this strategy to happen, because I personally think it's a pretty great and ingenious way to build some pretty important critical thinking, reasoning and relationship building skills.

See attached.

-Nick

[Mr. Mike]

This strategy is a great exercise in critical thinking but what examples can we come up with in real life that matches this situation?

[billylo]

I will keep this brief.

1. I like the current selection approach.

2. Like Frank, I am troubled by the need for teams who want the best chance of winning, resorting to select someone without intending to partner with them. This is not a behaviour that FIRST would like to promote.

To satisfy both needs, while maintaining the overall selection structure, I would like to hear your thoughts on a proposed adjustment to get rid of those awkward declines.

"The "Host your own party" card"

This 1 time only card can be used by the current team selecting its partners. Once used, all following team captains must "Host their own party" and be captains only.

This CARD essentially achieves the same effect, enabling the teams to enjoy their hard earned rights to break up the top teams, without going through awkward declines. As a bonus, it speeds up the process and remove the negative predicament of employing this strategy.

Thoughts?

[Pault]

Originally Posted by billylo

I will keep this brief.

1. I like the current selection approach.

2. Like Frank, I am troubled by the need for teams who want the best chance of winning, resorting to select someone without intending to partner with them. This is not a behaviour that FIRST would like to promote.

To satisfy both needs, while maintaining the overall selection structure, I would like to hear your thoughts on a proposed adjustment to get rid of those awkward declines.

"The "Host your own party" card"

This 1 time only card can be used by the current team selecting its partners. Once used, all following team captains must "Host their own party" and be captains only.

This CARD essentially achieves the same effect, enabling the teams to enjoy their hard earned rights to break up the top teams, without going through awkward declines. As a bonus, it speeds up the process and remove the negative predicament of employing this strategy.

Thoughts?

This wouldn't always be used, because as was mentioned before a lot of the time teams don't intend to use scorched earth, it just kind of ends up that way. And even if a team intentionally wants to use scorched earth, it could be to their advantage to pick a few teams first, in hopes that one would accept. Then, once they felt it was time, they would use the card and pick someone outside the top 8.

Anyways, I see no problem with scorched earth. I'm not going to list out my reasons why, because other people in this thread have already done that for me.

[Karthik]

I think it's neat the Frank sought out the community's opinion on this one. I also think it's neat that the community (at least the people vocal enough to post on ChiefDelphi) has pretty much come out an said they like the current system. Consensus is a good a thing. Of course, this is easy for me to say since I agree with the consensus that is being presented.

There's no reason for me to rehash why I think the current system works, since that's been taken care of quite eloquently by many of the above posters. But I do want to draw some attention to one major thing that can be improved. The current system works best when the best robots are ranked high. What can FRC do to ensure that the top robots end up being the top ranked teams?

Maximize the number of matches each team gets at an event.

There will always be randomness in FRC, because of the nature of the random alliances involved in qualification matches. The alliance system is central to FRC, and is not something that's going away. So to minimize the randomness it introduces, we need to play as many matches as possible. This is happening in the regions that have gone to districts, as well as Ontario, where teams are consistently getting 12 qualifying matches at events of 30-40 teams. This is great. Only getting 8 qualifying matches per team at the Championship in 2013 was simply unacceptable. The good news is that FIRST recognized this problem and is making efforts to fix it for 2014. Still, as FRC grows there will be a tendency to want to add teams to the Championship Event. This is fine, but FIRST needs to make sure that as they allow for growth of the event, they simultaneously allow for growth in the number of matches.

[D.Allred]

Originally Posted by Mr. Mike

This strategy is a great exercise in critical thinking but what examples can we come up with in real life that matches this situation?

Scorched earth strategy of a robot tournament doesn’t directly apply to my ‘real world’ experience. But I use this type of critical thinking everyday.

I have participated in technical qualification bids where my company competes against 6 to 8 other companies to win a long term agreement. You must understand your strengths and your competitors as they stack up the client’s requirements. Do you bid? Do you partner? Do you submit on your own strengths? 3, 2, 1, go!

FIRST has a recent example of this type bid where they requested proposals for the 2015 to 2019 FRC control system.

Competition is very real. Ethics is critical. Mentor wisely.

[Akash Rastogi]

Though not close to being the same type of situation, similar thought processes can be applied to game theory. Off the top of my head, similar styles of thinking are involved in the "Prisoner's dilemma"

I guess that's sort of "real world"

I just find it funny how many people think that this happens on purpose. Half the time "scorched Earth" happens is when an unexpected team ranks high and doesn't really know what to do so they pick down the rankings.

[Cory]

Originally Posted by Akash Rastogi

Though not close to being the same type of situation, similar thought processes can be applied to game theory. Off the top of my head, similar styles of thinking are involved in the "Prisoner's dilemma" https://en.wikipedia.org/wiki/Prisoner's_dilemma

I guess that's sort of "real world"

I just find it funny how many people think that this happens on purpose. Half the time "scorched Earth" happens is when an unexpected team ranks high and doesn't really know what to do so they pick down the rankings.

People have also blown this out of proportion both on Chief and in response to Frank's blog post. You'd think this is happening at every event and it's the only thing anyone is talking about-not true. There have only been two notable occurrences in FIRST history (at least that I can think of).

This rarely goes down as a true "pick from 3-8 and make them all say no". The only instance I can even recall something close to that happening was Curie this year. Far more often it means strategically picking 1-2 teams (that you would still like to play with!), causing them not to be ably to ally with the best robots when they deem you unworthy and decline, and then picking the best team that you know will say yes to you. It would be irresponsible not to do this if you knew you could not form a strong alliance as the number one seed.

I hope that Frank/FIRST largely ignore the specific advice given in the blog comments, because frankly most suggestions are horrible and will open the floodgates for teams to engage in legitimately nefarious behavior.

[DampRobot]

Originally Posted by Cory

This rarely goes down as a true "pick from 3-8 and make them all say no". The only instance I can even recall something close to that happening was Curie this year. Far more often it means strategically picking 1-2 teams (that you would still like to play with!), causing them not to be ably to ally with the best robots when they deem you unworthy and decline, and then picking the best team that you know will say yes to you. It would be irresponsible not to do this if you knew you could not form a strong alliance as the number one seed.

This reminds me of the Cookie's alliance selection play this year. If I recall correctly, they only picked 971 and 1662 and "made" them decline. It was definitely a good strategic play, that combined with them picking 973 was almost certainly the reason they made it to the finals.

Stuff like preventing the top 8 from picking each other or messing with how declines work will do much more harm than good, like you said. To echo other's sentiments, the real "solution" for this is just to play more qual matches.

[PVCpirate]

I'm not sure if anyone else has looked, but the discussion on the blog post comments is far more negative than this one.

[Jay O'Donnell]

Originally Posted by PVCpirate

I'm not sure if anyone else has looked, but the discussion on the blog post comments is far more negative than this one.

Frank noticed, and Frank became more awesome.

[Racer26]

I definitely agree with everyone. Scorched earth as a strategy is an artifact of teams doing what is in their best interest given their current position at the time of alliance selections.

If the seeding system did a better job of sorting the best robots to the top by playing more matches, and/or using sort metrics that are solely (or at least mostly) based on robot performance (ie. no 2012 Coop Bridges, 2010 QS), then you won't see scorched earth play out as often, because the team doing the selecting will be better than the team(s) they've selected.

Karthik is totally correct. Ontario and the Districts have done a good job. 12 matches in 30-40 team events seems to do a pretty good job of sorting based on WLT record. Its much harder to pull that off at Championship, without a huge change in how Championship works. To me, the obvious choice to both increase the Championship's team capacity, AND give every team more matches, without dramatically increasing time required, or volunteers required, is to move the divisions to having 2 fields each.

[Billfred]

Originally Posted by billylo

I will keep this brief.

1. I like the current selection approach.

2. Like Frank, I am troubled by the need for teams who want the best chance of winning, resorting to select someone without intending to partner with them. This is not a behaviour that FIRST would like to promote.

To satisfy both needs, while maintaining the overall selection structure, I would like to hear your thoughts on a proposed adjustment to get rid of those awkward declines.

"The "Host your own party" card"

This 1 time only card can be used by the current team selecting its partners. Once used, all following team captains must "Host their own party" and be captains only.

This CARD essentially achieves the same effect, enabling the teams to enjoy their hard earned rights to break up the top teams, without going through awkward declines. As a bonus, it speeds up the process and remove the negative predicament of employing this strategy.

Thoughts?

My thought: We have enough trouble explaining alliance selection to the crowd as it is. A card only adds further complications for which we already have a fix (read: just do the invites).

And if you don't intend to partner with someone, don't pick them. Remember, scorched earth can tank if an unanticipated acceptance crops up

[waialua359]

Originally Posted by Cory

People have also blown this out of proportion both on Chief and in response to Frank's blog post. You'd think this is happening at every event and it's the only thing anyone is talking about-not true. There have only been two notable occurrences in FIRST history (at least that I can think of).

This rarely goes down as a true "pick from 3-8 and make them all say no". The only instance I can even recall something close to that happening was Curie this year. Far more often it means strategically picking 1-2 teams (that you would still like to play with!), causing them not to be ably to ally with the best robots when they deem you unworthy and decline, and then picking the best team that you know will say yes to you. It would be irresponsible not to do this if you knew you could not form a strong alliance as the number one seed.

I hope that Frank/FIRST largely ignore the specific advice given in the blog comments, because frankly most suggestions are horrible and will open the floodgates for teams to engage in legitimately nefarious behavior.

On Curie this year, we were the 4th team to have declined 1678.
While I cant speak for the other 3 teams selected/declined ahead of us, I can tell you that it was totally unexpected. In hindsight, maybe we should have accepted? Maybe not?
We didnt scout enough about team 1678, but instead figured to select our own alliance with several teams we had in mind. Having to choose 2nd in the 2nd round was enticing to us as we looked for several key teams we thought would be there to complement our own.
We never had a discussion with 1678 at all during CMP.
However, it definitely was scorched earth in the sense that teams ahead of us definitely had others they wanted to choose instead, if possible.