Beta Bot Meta Broken?

FIRST General Forum
1

The question is: Does building a beta bot make sense every season even for top teams?

It seems beta bots are an integration test where you build a full robot. Or some think of it as a comp bot you build fast. Or even a more detailed Alpha bot to tide you over while the comp bot is getting made. If your beta bot’s a comp bot, is it too fast, is it provoking missteps? And if it’s an integration test, would smaller integration tests be better?

Similar to how you wouldn’t build an elevator as a prototype especially when you know how to build an elevator, does it make more sense to instead spend beta bot build time to do smaller integration tests of only the things you’re uncertain of/would have too much trouble fixing if it didn’t work well? Then after you’ve learned more and have higher confidence iterate on a comp bot.

Beta bots can suck your very experienced students into CADing and manufacturing a robot during time that could be spent running tests considering if you’re building the right thing. Often times pushing towards a fast beta bot means you didn’t consider details well enough so you’re just going to build many parts again, but slightly better (a manufacturing inefficiency), or scrap the design (a design and manufacturing inefficiency even if you did learn something, it could have probably been learned more efficiently).

I’m looking for thoughts on what a beta bot is to you and what you think a good way of thinking about it is - or any thoughts at all. How does the team and/or game affect whether you think it’s worth it to build a beta bot. For example, this year was it worth building a beta bot for your team? How do you know? Is there a year you are the most certain a beta bot was/would have been worth it?

More ramblings and thoughts

Part of this is coming from lots of rebuilds this past season. We saw a lot of high quality work from top and up and coming teams, but maybe not as much awe inspiring work. Is this a result of the games, I don’t think so.

If you do indeed see very large benefits from having a robot around during the beta bot period, should teams just be building simpler for the beta bot and leave spots for tacking other systems on later?

Alpha bots have seemed very useful. Generally I think week 1, anything you can build and know you are definitely making forward progress is a big win (drivebase, alphabot for programming/possibly driveteam, master sketches, prototypes…).

However, we’ve tried building a beta bot under good circumstances in 2020, and with a not as experienced team in 2024. In 2020, we didn’t learn as much as I would have hoped for one reason or another, and in 2024 we didn’t complete a beta bot well enough for it to be useful.

Purely using this as a conceptual example:
Because 1678 in my mind is tied to the beta bot concept, I can’t help but thinking of their main successes. Einstein 10 times running (seems like their expertise and ability to move quickly helps a lot here), and I think 1678’s main brilliancy, going for 5 stacks and focusing on can grabbers in 2015. The natural inclination is to attribute the 10 Einstein appearances to beta bot like concepts, and the 1 champs win, to a deep understanding of the game.

The question is,

  1. would it help for top teams to spend more time looking for brilliancies in design (I’m thinking 254 2018, 1114 in 2008, etc…) and
  2. If they were to take more time, even if just a little, looking for brilliancies considering mastersketches/crayon CAD/prototyping would this actually help or just negatively impact early season performance?

Have we tipped too far in the direction of move fast? Or do the learnings on finishing earlier, and testing with an actual robot provide more value than using smaller scale tests and thinking more, but more slowly.

If the ability of your team is lower, is the answer build a simpler robot and finish early allowing for iteration and practice, or does this have the potential to lead to missteps too often and you should instead build a simpler robot with more smaller scale testing, but less robot time.

5 Likes
2

I think the effectiveness of a beta bot will depend on team specific factors. Mainly how many students you have and how quickly they can iterate. Its up to you to decide if it is advantageous for your team to build one.

The most important thing at being good at anything, is experience. Building beta bots gives you twice as much experience at building a robot for that years game as anyone else. However whether or not you can harness that experience in time is the important part. 99% of teams are better off gaining experience building a robot in the offseason with no pressure perfecting a commonly used mechanism like an elevator that can be copy pasted onto any robot. The vast majority of teams don’t have the resources to perfect a sub-optimal robot in 7-8 weeks much less a theoretically perfect one in 6 weeks.

Teams that show up being perfect at one thing only in my experience always do far better than a team that can barely do anything but have the perfect robot architecture. in 2023 a team showed up week 2 and the only thing they had working was their drive train. but they ranked third. they were the only team that got on the charge station in auto and got the charge station RP every match because they stopped working on anything else on their robot come week 7 and perfected their balance auto.

4 Likes
3

It fully depends on the goal of the team. If you plan to have an Einstein level robot then building 1 is not a very smart plan since you should have at least double of everything.It would be a waste of resources to not run more than 1 bot for more practice time, tuning, trouble shooting and repair practice.

4

As a lower ability team (compared to the powerhouses listed), in my opinion this is the only way to go. Doing small scale testing is fine, but building a fleshed out set of prototypes, or a more complicated bot would take all our time, leading to our first competition being essentially testing and fixing things. Robot drive time (work out mechanical issues, programming testing, auton development) is the most important thing to have.

5

We had a spare of all our systems besides our drive base, but even then we had spare rails and modules. You don’t need to build a second robot to have a spare, I would even argue that having a 2nd robot doesn’t even qualify as spares.

I will say, we tried to do a beta bot (clone of alpha) in 2022 after doing it since around 2014. We ended up stopping, and the beta ended up nothing more than a chassis. We have since been able to accomplish wildly more complex and successful robots.

In terms of a simple bot to test out the game definition, I have wanted to do that for a while, but even then that’s mostly completed in prototypes.

1 Like
6

The trick is to build a Beta Bot with a Comp Bot build after it on the schedule, but build the Beta Bot to a high enough quality that you can ditch the Comp Bot build and just enter the well tuned Beta Bot when the schedule dies.

6 Likes
7

Can you define what a beta bot is?

1 Like
8

(Not 100% sure if you were responding to someone in particular Adam)

Now, I still 100% think there is value in discussing definitions here.

I have worked under the nomenclature of “proto bot”, which most closely resembles “integration test” in the OP. Mostly used for learning how to do new things and not worry about aesthetics when things need ti change (see welded powder coated frame problems). I think we learned things as a whole system, I don’t think smaller tests would have been as helpful as the full bot.

9

Yeah, good point, a little fuzzy on this. What im thinking of as a beta bot is a full robot integration test that also acts as a platform to experience as a whole (by driveteam, programming, mechanical, strategy etc…). With an emphasis on finishing early. The idea maybe being you have another opportunity to refine things after the beta bot, by rebuilding or simply refining.

10

Our team is planning on building a swerve kitbot for a purpose similar to what you describe. But it is also to keep the large amount of rookies we have busy and learning hands with full robot experience.

11
4 Likes