Was "Sandstorm" the best addition to FRC since bumpers?

This was a bit of a clickbait title, but I think pretty much everyone agrees that the addition of bumpers was a great addition to the “evergreen” items that make up just about every standard FRC game since it’s inception (ignoring you 2015).

I propose that the modification of the autonomous mode to the “sandstorm” mode this year might have been the best thing that FIRST has done to the standard model of an FRC match since they added bumpers.

The sandstorm change raised the floor of what the lower and mid level of teams were able to accomplish during those 15 seconds, and the amount of times where 4-6 robots either sat still or only drove forward was noticeably much lower.

The top teams performing at the highest level still had autonomous routines as having something like a 2 hatch routine was much more reliable then driving it. I also think as a by-product of teams not feeling as much pressure to get an autonomous routine done, more teams had working vision tracking software (definitely also a by-product of the limelight) and just overall better performance in teleop period.

The biggest drawbacks that I have heard people complain about overall has been a removal of a programming task. I think overall most of the teams that in the past have taken advantage of high-level autonomous modes still had one this year. I think for the rest of the teams out there, there are many programming tasks within just the control of the robot for teleop that could use a lot of refinement and improvement that would still benefit students education just as much or even more then throwing together some mediocre autonomous routine.

I’m curious what the overall opinion of sandstorm was this season, and if people think it should be something that FIRST continues to do in future FRC games.

  • I liked Sandstorm, would like to see it in future games
  • I liked Sandstorm, would not like to see it in future games
  • I liked Sandstorm, not sure if I’d like to see it in future games
  • I’m undecided on sandstorm so far
  • I disliked sandstorm, would not like to see it in future games
  • I disliked sandstorm, would like to see it in future games
  • I disliked sandstorm, not sure if I’d like to see it in future games

0 voters

9 Likes

The biggest part of Sandstorm I liked was the de facto “A Stop” functionality it provided teams. We’re an ambitious team when it comes to our autonomous goals, but I do not allow my team to test new or partially proven routines on the field (especially if we have other proven routines we can count on). Anyone who witnessed the amount of time we spend on the practice field validating paths and routines can attest to that. Having that A Stop functionality enabled me to be a little less stringent about running routines that may not be 100% proven yet, since I knew we could abort any routine that wasn’t going well (and even take driver control to try and salvage it).

That being said, if there was once again a ranking point tied to autonomous (as there was in 2018), even an A Stop wouldn’t make me particularly flexible about running less consistent routines.

11 Likes

biggest gripe I have is cameras, and FMS bandwidth. But there were only 2 particularly bad matches that I saw that had a noticeable effect on the gameplay so maybe its not that bad. I am interested to hear peoples opinions on having no blinds. Essentially a 15 second time period where you cant cross the field and play defence or maybe you cant leave the corner of the field you started in. or some other type of handicap or interference that makes it still advantageous to run an autonomous.

8 Likes

I loved it because even teams that don’t have strong programmers could get by with a camera and some driver practice. Things that raise the floor like this are always a big win in my book. I hope they stick with it next year in some way.

6 Likes

I loved it. It provides a new driver challenge of driving in first person view and still awards true autonomous modes. It also rewards semi-autonomous modes such as a team driving close to the rocket and then starting an autonomous “score” routine and then driving back to the loading stationand starting an autonomous “retrieve” routine. Those semi autonomous routines directly translate teleop whereas most autonomous routines see no use once the first 15 secondsof the match are over.

The one disadvantage I saw is that many teams would say “Lets just plop a camera on and let the driver wing it. No need for the programmers to get the robot to do any autonomous programming” That could be disheartening to programmers who are used to doing autonomous routines. But again, it could just have programmers spend their time improving teleop actions (more closed loop, semi-automated actions, etc)

That was a rant…but overall I really liked it. I hope it wasnt but also fear it was just this year’s game-specific gimmick and we won’t see it again. If we do I look forward to the convoluted ways the game calls for a blind period.

2 Likes

TBA data shows around a 3% increase in crossing the line in sandstorm during all qual matches at 90.1% compared to 87% in 2018 and 87% in 2017. With around a 1% increase in the playoffs compared to 2017 and 2018. IMO this is not worth removing the autonomous mode as it leads to teams being stuck in the middle where they are good enough to drive off the camera and score but will never take the leap and try to make an auto that can compete with top level teams. They will also miss the learning opportunity and the cool factor of something that is fully autonomous. It was nice having the ability to stop an auto if things were going wrong though.

51 Likes

All i will add is that with it being a city theme and the game name appearing to do with electricity, I think it’s very likely we see Sandstorm return next season, but renamed “blackout”

8 Likes

Any chance we could grab data on game objects scored in 2017 or 2018 versus 2019? I think if you are going to compare data points “cross the line” shouldn’t be the only metric being looked at here.

2 Likes

2019 doesn’t have anything specific about sandstorm scoring since it wasn’t worth anything extra so we’re not able to compare it that way unfortunately. Would you be in favor of no sandstorm or auto and just have teleop last longer? Maybe keep the can’t cross the line for 15 seconds so teams can still do auto if they want to? Or is having the curtain needed to make it so that autos if done well can be better than drivers?

1 Like

I think the latter is what makes it a pretty good balance. A couple other potential solutions I have come up with so far:

  1. An autonomous “skills challenge” similar to what VEX has that maybe is implemented into the rankings somehow. Logistical challenges here are how do events create enough time to do this and how is it integrated into the rankings?

  2. Keeping sandstorm pretty much as is for qualification matches, and then making the period autonomous-only (maybe with an autonomous e-stop button) for elimination matches. This way the floor for the lowest team is still raised for them to be able to contribute in qualification matches, while the majority of teams who are aiming to do well in eliminations still have a lot of incentive to develop autonomous routines. It also presents some strategy on whether you run your autonomous routines in qualification matches to show scouts, etc. versus driving in sandstorm if your driver performs better then your routine.

I think #2 might be my favorite solution. Would need to figure out logistically how the FMS regulates that difference between quals and elims, but I think this keeps the benefits of raising the floor overall while still having the challenge of doing auto routines and them still being very incentivized.

2 Likes

I liked it and hope to see something similar in the future.

Our team treated it just like autonomous mode and that worked well for us. I don’t think Sandstorm should have sent most teams into thinking “no programming” right away, which was the biggest complaint when it was announced…no need for programming.
Clearly good autonomous modes are likely faster than a driver with a camera (which was our reasoning), but for those teams who have lower programming skills, I can see why they went that route. (and some got pretty good at driving during that time!)
Having this time be a choice for the team to do driver or autonomous was nice though, especially if an auto failed and then the driver could “bail” and take over.

Issues were the need for a camera. Field connectivity with a Limelight seemed to cause problems for many. Laggy systems, FMS bandwidth, etc. So huge drawback there that if Sandstorm continues to be a thing would hopefully be resolved somewhat.

Nothing extra point wise for this period. Sure you could get ahead scoring wise, but no extra ranking points like in previous games, which in my opinion was a slight drawback for those that went farther than just crossing the line.

1 Like

I liked sandstorm, but this is similar to what I’m afraid of. Providing a way for teams to start down the path to autonomous mode is great, but it could lead to teams not focusing on auto as much as the reward is much lower.

P.S. I preferred no bumpers, so sandstorm was far superior

Unpopular opinion here (at least based on the poll results so far): I thought sandstorm was a terrible addition.

As someone who, as a student, almost exclusively did programming, strategy/scouting, and awards, there are pros and cons here:

Pros:

  • Programming takes less time, so now student Jeff can focus on awards, scouting, and strategy! Yay!
  • Now we don’t have an excuse to not have a camera at our first event/use a camera to assist drivers

Cons:

  • Autonomous is really where programmers get to shine. Sure, the robot would not move without programming, but there seems to be (at least when I was a student) a culture that dictates drive team and fabrication are the “best” parts of the team because that’s what’s seen on the field. Autonomous period really felt like my little moment (or quarter of a moment I guess) where I could say, “Yeah, I did that.”
  • To quote Kaleb here:

to which I wholeheartedly agree.

  • Darude Sandstorm. Need I say more?
  • When I was a student, we always started our robot against the back wall to make sure our distances were correct should we run an auton. If there has to be a curtain for sandstorm in the future, that certainty is gone (granted, I guess most teams would go with driver control for this)

Now, as a new (albeit not totally active) mentor primarily focused on helping the programming subteam, I felt like I wouldn’t be needed after hearing the announcement of Sandstorm. For the most part, the students knew how to code the teleoperated period, and Sandstorm is basically teleoperated with a camera. What am I supposed to teach the students, how to look up two lines of code?

If Sandstorm begins showing up in future games, I will honestly be pretty disappointed. Even as a mentor, I still notice that programming subteams aren’t as recognized as other subteams. Additionally, lower-tier teams can typically find help from higher-tier teams in at least getting a line-cross auto coded. Now, with Sandstorm, there’s less incentive for the lower-tier teams to seek out that help, and less incentive for the higher-tier teams to reach out to the lower-tier teams, which in itself creates plenty of positive side effects from the student-to-student interaction between teams. With Sandstorm, I could see a “well any team can just drive forward a few inches and they won’t get in our way, we don’t need to check in with them” (which anybody in FRC for any length of time knows is completely false) taking over.

Tl;dr I didn’t like Sandstorm because it fueled the culture problem surrounding the perspective of programming subteams, decreases student-to-student interaction between teams, and disincentivizes mid-tier teams from pursuing auto, even if it is well within their capabilities.

11 Likes

I don’t really agree here. It’s not like programmers sit on their hands and call it a day. The automation of routines like scoring and intaking are still critical tasks a team can take on when they don’t have to add in things like path planning.

The way I see it: teams were encouraged to learn how to get cameras added which is a huge plus. They could then focus on automation tasks other than getting a gyro and encoders to get them to a specific location.

6 Likes

There are also other “hybrid” sandstorm/autonomous options, especially if you are okay with enforcement by referees rather than FMS. Things like only one team per alliance isn’t allowed to touch controls during the sandstorm, but the others can. Or maybe there are five seconds where nobody can touch controls, and then 10 where they can, and then the curtain lifts.
Obviously there are disadvantages to making it more complicated as well.

1 Like

I guess it would depend on the team. We have absolute no programming skills so even if we did have our own programmer, i would feel like we’d only do the bare minimum (see 2018). From the time I’ve been on the team (2017), we’ve only went forward, put the gear on the peg and wait 10 seconds and back off into teleop. 2018 was just a cross the baseline. Teams with a good base are those who really succeed.

Sandstorm was a mixed bag this season.
On one hand:

  • More teams were able to reasonably contribute during the first 15 seconds of the match
  • Teams were able to A-stop if their auto routines were ineffective
  • There were still legitimate reasons to use autonomous or automatically assisting functions to make advanced autonomous routines more consistent
  • Teams across FRC implemented cameras that might not otherwise have

On the other hand:

  • Teams had little reason to implement advanced autonomous pathing, advanced vision systems, and other programming challenges that were both a great learning experience for students and a raised ceiling for robot ability - multiple teams managed to hit two hatch sandstorm routines without any autonomous capability.

I think a good compromise between these two in the future would be to allow teams to drive during autonomous, but if a team implements an actually autonomous routine, their autonomous could be worth more points - perhaps game pieces scored autonomously would have a bonus value attached to them, but if the drive team touches the controls, the game pieces are worth the normal amount.

5 Likes

2791 had an elite-tier Sandstorm mode that was purely driver control.

I want there to be a Sandstorm like mode in the future, but eventually teams are going to catch on and realize that in a lot of games, running autonomous modes in Sandstorm will slow you down if you aren’t 254-tier in your software.

That said, I highly value that teams who can’t do advanced programming aren’t just hobbled from the start and unable to play at the highest levels. Hopefully there’s a balance, or some nominal point bonus for not touching your controls in a match.

Don’t get rid of A-Stop tho.

5 Likes

My students were not a fan of the Sandstorm period. This was our 8th
year and last year we really pushed our programming team to work
more on autonomous. Our number one programmer responded and
by the end we could perform all of the tasks.
My question each year to the students is “do we build a robot”. Their
response is mixed, because we really build a remote control vehicle.
One of the challenges must be automating any function, this moves the
device closer to a robot.
I know this is a challenge for smaller or newer teams, but there must be
something to strive for.
Just an opinion

1 Like

I liked and disliked sandstorm at the same time.

One of the bad things about sandstorm is that a programmer can program autonomous modes and just not have them tested.

I programmed a few autonomous modes. They weren’t the best thing in the world, but if we were forced to use them, they would have improved over time. With the sandstorm, teams like us aren’t encouraged to refine autonomous over the course of the season, which sucks.

2 Likes