1678 Two Ball Auto
 

After a lot of work, we're proud to present our two ball auto routine!

https://goo.gl/photos/NLA2oktnyhnwoRjV7

Re: 1678 Two Ball Auto
 

I can't wait to see this happen in a match. So far this has seemed impossible. I'm optimistic that it isn't impossible.

Re: 1678 Two Ball Auto
 

Welp time to program this for our team. Nice job with your two ball auto!

Re: 1678 Two Ball Auto
 

Congratulations!!!

I was almost sure this would never happen. The depositing of the first ball is key here.

In respect of the Loch Ness Monster stickers of 2012, perhaps the 2-ball auto teams should get a big fat bench grinder sticker to place on their robot!

Re: 1678 Two Ball Auto
 

Thats insane Nice job guys

Re: 1678 Two Ball Auto
 

That's a nice bench grinder

Re: 1678 Two Ball Auto
 

Wish people would post pictures of their robot, for all I know this is a benchtop grinder shooting the balls

Re: 1678 Two Ball Auto
 

#TheLittleBenchGrinderThatCould
#TwoBallOrBust

Re: 1678 Two Ball Auto
 

Oh neat! You guys initiated the challenge this year ;)

Re: 1678 Two Ball Auto
 

Good to see we weren't the only ones with this idea. I had a feeling 1678 would be the ones to beat us to it! :D

Re: 1678 Two Ball Auto
 

While dropping the 1st ball helps in preventing another team from trying to steal or move the ball on the ball on the center line, it is not key to this auto.

The placement of that boulder on the center line is in a location that could hinder the other alliance from crossing a defense should they choose that, as the secret passage is directly in front and not a defense. So it comes down to do I stop them from scoring 10 points or do I score 10 points, also do I risk hitting them if we go for that ball at the same time?

Also awesome auto, though i don't think that bench grinder is a legal FRC robot.

Re: 1678 Two Ball Auto
 

Sigh... now our programmers are going to want us to build two practice bots too...

Well done. Not at all surprised that you guys made it happen.

Re: 1678 Two Ball Auto
 

Fantastic job as expected from 1678. Great job getting the timing down.

Re: 1678 Two Ball Auto
 

I don't think dropping the first ball is motivated by wanting to "steal" it from the opposing alliance. They're simply putting their starting ball in a place where they have significantly more leeway on the 2nd run without getting a penalty. Why risk trying to get within 2" of a spot after crossing defenses if you can be much more reliable doing it BEFORE you cross any defenses.

1678 looked at the problem, and found a way to solve it that should be significantly less risky. If they miss the 2nd ball on the way back, oh well, but at least they're much less likely to get a penalty while doing it. Pure genius, and major kudos.

P.S. How long until you can do 3? :-P

Re: 1678 Two Ball Auto
 

I still think the dropping of the first boulder is super key. If you shoot your starting boulder first, you run into a huge problem. How do you approach the midline with enough precision to grab that second boulder, but not get a penalty for crossing the midline in auto. That's a very tall order for a dead-reckoning or vision-tracking robot, which may not know where it is at all times.

Drop your starting boulder far away from the midline, and you don't have to be nearly as precise.

Re: 1678 Two Ball Auto
 

Quality bench grinder.

Re: 1678 Two Ball Auto
 

I think another interesting question is will we see this at the first event for some of the other top tier teams.

Re: 1678 Two Ball Auto
 

Very cool, however since your going over the same defense twice its not worth anymore than just doing the second ball in teleop, 30 points regardless, but it does save you time in teleop. kudos on making it work.

Re: 1678 Two Ball Auto
 

The second ball is scored as an autonomous ball, which is 10 points versus 5 for the same ball. Since you still have the ability to get the 5 points for the second crossing (that you didn't get in autonomous) later in the match, it is a net gain of 5 points across the whole match doing the same behaviors. It is definitely "worth more" than just doing it in teleop. 1678 wouldn't do things like this for no reason.

Re: 1678 Two Ball Auto
 

It's worth more both points-wise and time-wise. Alliances who can score a combined 3 balls in autonomous have to shoot a little more than half as many balls as they would have otherwise. Teams with two-ball autos will be able to get captures more frequently, rank higher, and have a higher chance to win.

Re: 1678 Two Ball Auto
 

Having seen this robot, It's kinda funny they chose a bench grinder. It's got the noise of one without the 5817-esque trajectory.

Re: 1678 Two Ball Auto
 

Huh. Well I'll be darned.

Re: 1678 Two Ball Auto
 

So glad to see you were able to reach the two ball auto goal. I knew it was rumored, now it's confirmed!

See ya'll in a couple days!

Re: 1678 Two Ball Auto
 

Not that this isn't impressive, but I'd call this more "plausible" than confirmed. Only when it's done during an official match would I give it confirmed.

Re: 1678 Two Ball Auto
 

I smell a turret. GG 1678, impressive yet again.

Re: 1678 Two Ball Auto
 

Really wish this was just a video of the robot and not of a bench grinder meme.

Re: 1678 Two Ball Auto
 

Wow!

Re: 1678 Two Ball Auto
 

I wouldn't really rely on your smell then. :o

Re: 1678 Two Ball Auto
 

Kudos for thinking outside the box. Dropping the original ball to pick up one off the midline while you know exactly where you and the ball are is really clever. I gave this problem some thought but didn't even consider this approach.

Re: 1678 Two Ball Auto
 

I don't know if this post inspired the approach, but Dr. Joe certainly got my attention with it back on January 26th.

By the way, well done 1678. Impressed yet again.

Re: 1678 Two Ball Auto
 

So either they don't have cloth hanging from their low bar, which I think could affect their timing and accuracy, or they're shooting with the low bar cloth hanging on their robot for the second shot. Which seems unlikely.

It's an impressive achievement, but there have been a lot of impressive auto videos that have mostly or completely failed on field.

Re: 1678 Two Ball Auto
 

Let's agree to disagree here.
Working with 1678 this summer is enough to convince me that they don't need Pixie Dust to make magic happen. The use of Pixie Dust remains firmly realm of 2073.

Re: 1678 Two Ball Auto
 

I'm glad 1678 proved it could be done. Now our programmers have no excuses!

Impressive, as would be expected from a reigning World Champion. Good luck 1678!

Re: 1678 Two Ball Auto
 

I was waiting for the first signs of this auton this season...

Congrats to 1678 for winning the completely arbitrary race for the 2 ball autonomous cookie for Stronghold!

After the past several seasons, you were definitely on the short list of teams I would have put money on for winning it. Best of luck this season, guys!

Re: 1678 Two Ball Auto
 

You must be real fun at Parties.

Re: 1678 Two Ball Auto
 

Cool auto mode, super impressive! I always look forward to seeing what 1678 cooks up. Can't wait to play with/against you at CVR!

Re: 1678 Two Ball Auto
 

Wait, there was a cookie on the line?

We would have totally posted earlier if we had known ;)

And thanks, our kids have worked crazy hard this season, I'm one of many proud mentors!

-Mike

Re: 1678 Two Ball Auto
 

Awesome 1678 can't wait to see it . I had that same move laid out week 1 of build season sadly I doubt we'll get it down. We need to get the HG down first :)

So excited to see all the word class teams at CV

Re: 1678 Two Ball Auto
 

Guys, you should know by now that it's best to operate under the assumption that a cookie is always on the line. ;)

You've got a lot to be proud of, and justifiably so. Hope to finally get the chance to share the field with you guys in St. Louis this year.

Re: 1678 Two Ball Auto
 

^This. Not to take anything away from 1678's autonomous, but doing it on the practice field is a lot different on an actual field (I think they would agree with this). It was only two years ago they posted a very similar video of their 3 ball auto and never attempted it at their first two regionals, citing consistency issues.

All that being said, it is still the most consistent 2 ball auto I've seen this year :cool:

Re: 1678 Two Ball Auto
 

The bench grinder was only a memes an end. We feel the need to protect our design and IP for as long as possible to maintain a small competitive edge.

It was either the bench grinder, a troll face, or our team logo. One was professional, one was decidedly in poor taste, one was a hilarious play on one of this year's memes. We were never going to reveal the robot itself until CVR, that has been our decision since before kickoff.

Re: 1678 Two Ball Auto
 

Thanks for the Gracious Professionalism.

I was noting some weird/strange about the video that no one else had commented on, and also that team defenses, fields, and balls are not equivalent to said items at the event. But your flawless logic has convinced me: there is no way this ever fails in competition.

Re: 1678 Two Ball Auto
 

I'm definitely going to parties with you two from now on.

Re: 1678 Two Ball Auto
 

Re: 1678 Two Ball Auto
 

That is a scary amount of GP you have there sir. And I've seen lots of GP in my days.

Re: 1678 Two Ball Auto
 

Did you accidently get sand in your shoes and not just in the bag? ;)

Re: 1678 Two Ball Auto
 

I have first hand experience in knowing how true this statement is. Just because an auto works 10/10 times in a team's own practice field (or even during a practice match at the event) does not mean that it will work during competition. That being said there is no doubt that 1678 will make a super consistent 2 ball auto happen in an actual event (maybe not immediately, but it will still happen).

Re: 1678 Two Ball Auto
 

Even if 1678 has some trouble getting this going in a competition (I'm sure they will nail it eventually), posting this video has raised the bar for all teams attempting to score big in autonomous mode. I don't doubt seeing this video will inspire some teams to go out and spend a bunch of time making their autonomous routines better (my team included). Stronghold will be a better game for everyone to watch and play with robots doing more in autonomous mode.

I don't understand why some people are putting down 1678 in this thread. I love that they posted this video.

Re: 1678 Two Ball Auto
 

Agreed I know 1678 will find a way...might not be till end or could be match 1. The field IS different and all teams will find that out in a hurry their first match. Can't wait to see it and prove to our team "I told you so". Literally day 1 I had it all planned out. In fact we designed our intake past the bumpers for this very reason 30-point auto to suck in from midline. Who knows maybe we'll find a way to get auto aim going in CV. The code works its just camera placement (to avoid destruction of it) and real estate on the low bot. At this point I'd be ecstatic with a 20-point auto. Please auto-aim revelation come to us by CV (That what worked for our scaler)

Awesome solution to the mid-line potential foul...Brilliant. CV is going to be a blast. Hope you nail it 1678. Wow.

My original plan was for us to start in spy and go from there....thanks to Team 3316 for the shooting chart (awesome job)

http://imgur.com/bYqeIIs

Re: 1678 Two Ball Auto
 

This video if anything, shows how most of the rest of the powerhouses plan to do it. This especially in addressing how teams plan to acquire the 2nd ball without crossing the line.

Re: 1678 Two Ball Auto
 

Wow. Thanks guys. Now I have to listen to Brandon ask why we aren't doing this.

For real though, this is just awesome. Looking forward to seeing y'all compete.

Great job and good luck!

Re: 1678 Two Ball Auto
 

For posting this and letting the genie out of the bottle :(

Re: 1678 Two Ball Auto
 

The reaction in this thread is very similar to the reaction in other threads where teams post teaser videos but hide their design. Not saying it is fair or justified, but it is consistent with how CD responds.

Re: 1678 Two Ball Auto
 

Agreed. I see a lot of "If there ain't no video, it didn't happen" type responses.
Clearly there is video. Clearly it is possible. It just is NOT EASY!! Otherwise we all would be doing it.

Re: 1678 Two Ball Auto
 

Now I thought my teams auto was great...but this just blows ours out of the water!!!

Great job guys!

Re: 1678 Two Ball Auto
 

Wesley who posted this video was the freshman hiding in our 2014 reveal video showing us throwing in a 3-ball auto! :D

Re: 1678 Two Ball Auto
 

Shh Richard don't give away our secret sauce.

Re: 1678 Two Ball Auto
 

Do you guys have a name for this robot yet? If not, we're calling it "The Grinder" from here on out!

Re: 1678 Two Ball Auto
 

This is amazing, kudos to you guys for yet again another amazing achievement. This is just another reason to my programmers are gonna hate me now!

Re: 1678 Two Ball Auto
 

Our 2016 robot is named "Adrian" in honor of our beloved 1678 Alumni and Mentor Adrian Clark, who passed away unexpectedly this past summer.

We are humbled to carry on his legacy through a program he loved so much.

-Mike

Re: 1678 Two Ball Auto
 

We have named him Adrian in honor of one of our young alumni mentors that passed away in June of last year. He was the one responsible for solving our biggest consistency issues with last years can grabbers the day before we packed up for Champs. He was also the mentor I worked along side of the most working with our Mechanical Design and Fabrication sub-teams last year. The last time I saw him was when we wrapped up after picking up and unloading the 5 big machines we got a couple of days earlier. I am sad that I wasn't able to work with him this season on improving our shop and training our students. He will be greatly missed.


EDIT: also the stuff Mike said
-------------------------------------------------------------------------------------

When we started working on the 2 ball auto we were getting times of around 23 seconds for the full cycle. We managed to shave 9 seconds off of that time(And we could have done more!!!)*

*well probably not any meaningful amount of time

Re: 1678 Two Ball Auto
 

Will you guys be releasing an unedited version of this video after your first competition?

Re: 1678 Two Ball Auto
 

This. We have Utah the same as CVR so I won't be able to watch it live.
As a side note I find it a little ironic that the same people who condemmned 5817 for hiding a shooter prototype early-season are hiding their robot bench grinder now, when it's too late to make significant design changes. An unedited version after CVR would be neat.

Re: 1678 Two Ball Auto
 

Jesus christ.... please have mercy on the central valley....

Re: 1678 Two Ball Auto
 

Is there any reason you guys shoot the ball on the field first, and then use the ball you started with?

Re: 1678 Two Ball Auto
 

That's American GP.

Re: 1678 Two Ball Auto
 

me walking into this thread after being at work and meetings all day


Congrats to 1678 for managing to build a team that has developed a concentrated level of what is likely insanity-based-dedication across a wide group of people. Win world champs and you can take the first 15 weeks of the season at a slowly increasing pace instead of blistering, right? Guess that's not the attitude y'all have.

Re: 1678 Two Ball Auto
 

One ball the competitor, two balls the champion, three balls the dream?

Re: 1678 Two Ball Auto
 

Quoted relevant points from this thread.

Re: 1678 Two Ball Auto
 

Honorable name. Adrian it is!

Do well Adrian!

Re: 1678 Two Ball Auto
 

Now imagine if you had two robots like this on one alliance! 60 point auto, not even counting the third robot, so potentially 70 or 80 point. In addition, the 4 balls scored in auto and the minimum 3 balls starting in the castle means that the opposing HP is forced to release at least 1 ball at the beginning of tele-op. At that moment, all opposing robots will be on the other end of the field, allowing this ridiculously overpowered alliance to grab the boulders off the secret passage without interference. This alliance would have a HUGE lead in points and cycle time!:ahh:

Maybe I will to see this at IRI or on the Einstein field. One can hope...:]

To make this possible, one of the autonomous routines must be performed on a non-low bar defense. 1678, your move.;)

Re: 1678 Two Ball Auto
 

So what does that make 2 Champs?...

Re: 1678 Two Ball Auto
 

6 ball auto confirmed

Re: 1678 Two Ball Auto
 

Cross a Defense: Dukes of Easy mode
1 Ball Auto: Hard mode
2 Ball Auto: Harder mode
3 Ball Auto: Hardererest mode
4 Ball Auto: God Mode

I will bet a 2 inch Turner's Cube that we won't see a successful 3 ball auto all season in a match. A 4 inch Turner's Cube to the team that does a 4 ball in a match, and me wasting hours on figuring out how you hacked the very laws of physics.

Re: 1678 Two Ball Auto
 

It is definitely worth coming up with a 3 ball auto to get one of these, they are absolutely fantastic. Well worth the shirt I traded at Champs last year!

Re: 1678 Two Ball Auto
 

I kinda think that 254, 1671 and 973 won't be too intimidated...

Re: 1678 Two Ball Auto
 

Pretty sure low bar is a necessity to pull it off (talk about auto traffic jam) ..doubt two robots pull that move off consitently but who knows. I think the single boulder possession rule make this nearly impossible.

Re: 1678 Two Ball Auto
 

It seems possible (though extremely hard) to do it with any other defense, except maybe the ramparts. I doubt it will happen before Einstein, if at all, though.

Re: 1678 Two Ball Auto
 

Only thing that cancels ONE successful 2 ball auto (20 besides the cross) are 2 scaling bots (20 extra) so if 1678 pulls this off of the other alliances better find some scalers against them or 1678 better snatch up some scalers as captain..just sayin'

I think 1678 just made scaling bots top of peoples pick lists to defeat that alliance

20 points per match is 4 HG or 4 2x crosses or breach/capture...its huge

3 scalers are 1 HG better than 1 2-ball auto

Most everyone can cross in eliminations.

Re: 1678 Two Ball Auto
 

You (generally) want to maximize non-excludable points per robot per alliance in any case - these robots are no more or less valuable than they were before. You always wanted the points. I'm not picking a hanger specifically over an otherwise higher scoring shooter just because of 1678's auton (though this is admittedly a contrived situation).

Re: 1678 Two Ball Auto
 

However the 2 ball auto also lowers the tower by 2 points.
Plus whatever your alliance members put up in auto.
Which will make it easier in tele-op to achieve a capture.

Scaling might offset the two ball in terms of points, but a two ball auto still is the stronger strategy IMHO.

Re: 1678 Two Ball Auto
 

Two ball autos can be pretty easily blocked though, while scales can not.

Re: 1678 Two Ball Auto
 

Explain? Seeing as you can't cross the midline, it is pretty hard to stop a two ball auto. The only way you could do it is a canburglar-esque race to the center ball, but seeing as they immediately intake the midline ball without moving, it seems pretty unlikely that you will be blocked.

Re: 1678 Two Ball Auto
 

Well for one if you attempted to block an auton at all you would be on the wrong side of the field, which is illegal.

Re: 1678 Two Ball Auto
 

Just start without a ball and have your intake ready to grab the ball as soon as the match starts. 1678 has to drop their ball first, so it would be relatively straightforward to beat them to the center ball.

There are more complicated 2-ball autos that could avoid being blocked, but considering we went through week 1 without any 2-balls at all, we probably won't see hardly any two ball routines more advanced than 1678's.

Re: 1678 Two Ball Auto
 

Thanks great idea..almost like can wars fastest intake wins.... ours is really fast (Fastest at SD)

Re: 1678 Two Ball Auto
 

It still requires a specialized auto mode, where you're lined up against the secret passage rather than a defense. So either your entire auto mode is preventing a two ball auto (the 1-ball will still work), or you're coming up with a much more complicated (read: more likely to fail) auto mode.

Yes, you can TECHNICALLY defend a 2-ball auto, but I don't believe the tradeoffs are worth it.

Re: 1678 Two Ball Auto
 

Who says you can't still cross after doing this?

We plan to have such an automode ready for champs.

Re: 1678 Two Ball Auto
 

Nothing says you can't. It's just more complicated than it initially sounds if you're not thinking about field configuration.

Re: 1678 Two Ball Auto
 

Looks like we'll have to add our surgical tubing powered, 1" diameter carbon fiber arms to grab the ball at 120mph ;)

I'm getting flashbacks to 2015, the horror! :ahh:

-Mike

Re: 1678 Two Ball Auto
 

I am thinking about the field configuration, and it's not that complicated. Certainly easier than a 2 ball auto.

Re: 1678 Two Ball Auto
 

Didn't mean to imply you weren't or that it can't be done. It certainly can (and likely will).

The point I was trying to get across (poorly, perhaps) was that if you're blocking a two ball with your auto mode, but this prevents you from doing anything positive for your own alliance, then its not worth it. Having an auto mode that "steals" the center ball then still does something useful afterwards will likely be one of the more challenging routines.

Re: 1678 Two Ball Auto
 

The secret benefit to a 2-ball auto: if your partners also score their boulders in auto, your opponents are forced to stuff a ball into the secret passage right away. At Worlds, this will be the game changer. A tower won't stand a chance when it's only 3 damage away from a capture in the opening seconds of teleop.

Re: 1678 Two Ball Auto
 

Second best defense for 2-ball auto is the portcullis if you have a mechanism that lets you reverse through it. Or since you need a pickup on your back side to do this, probably a mechanism on your front side as well as on the pickup. After that your percentages probably drop by a lot since you're likely being randomized by the defense or the landing afterwards.

Re: 1678 Two Ball Auto
 

Oooooo I like the way this sounds, not as a defense against the 2 ball auto but as another position a 2 ball auto could take place in, I'm glad that not only did we decide to put a herder on the front and back of the bot, but both herders are capable of opening the portcullis. Time to push the programmers even harder!

Re: 1678 Two Ball Auto
 

OTOH, any defense that's not the low bar has increased difficulty because you don't have a wall to stop your dropped ball. That cardboard field border looks pretty integral to 1678's auto routine. Without it, you have to worry a lot more about an inconsistent ball running away from you.

Also, also, an alternate, option for interfering with this style of 2-ball auto: instead of stealing the first ball (maybe you're not fast enough), bump the second ball with your ball. This assumes you have nothing better to do with your ball, and that you can outfeed fast enough to get both balls clear of the infeed path. Still, second picks often have nothing better to do with their ball, and it'd be an easy routine to write. Oh. Bonus points if you steal the first ball, and then shoot it at the dropped ball to turn a 2-ball auto into a 0-ball auto.:cool:

Re: 1678 Two Ball Auto
 

FYI, We will be showing off the unedited version of the video (i.e. no grinder) on our show on Wednesday. Thank you so much to the students and mentors on 1678!

If you want to catch it live, check it out Wednesday at 9 PM ET at twitch.tv/FRCGamesense or youtube.com/FRCGameSense

Re: 1678 Two Ball Auto
 

This is really the biggest crux around auto this year. The more you score in auto (where you don't have to worry about defense, just run the same routine you've run all season), the less you have to score in teleop. Every ball your alliance scores in auto means just that much more time for every cycle, and gives you leeway should you run into a little bit of shot inconsistency or heavy defense.

Re: 1678 Two Ball Auto
 

I'll remember that.

Re: 1678 Two Ball Auto
 

Yes, you will be able to see it on GameSense this Wednesday.

Link here: http://www.chiefdelphi.com/forums/sh...40&postcount=2