Is Climb Roulette Back?

FIRST General Forum
1

Is it back? For those of us around for 2017, a term called “Climb Roulette” became common place. The general idea was that the end game points were weighted so heavily that if a team on your alliance failed to climb, it could torpedo your chances of winning a match, regardless of what else you do during the match. Steamworks had some really poorly balanced scoring all around, but especially the climbs were valued egregiously high.

This year a climb is worth 25 points. Compare that to a “maximum cycle” being worth 15 points, and a “realistic cycle” being worth 10 points, and a climb is worth 1.5-2.5 cycles. The odds that you’re able to complete that number of cycles in the same amount of time it takes to climb is extremely low. This combined with the fact that there is plenty of room for 3 robots to climb means that climbing is strategically advantageous for every team.

Climbing is old news. It’s a solved problem. The added twist of balancing is certainly interesting, and will enable some teams to provide themselves with an advantage through shifting their weight, but most teams should be able to do a simple pull up.

This is all to say that there should be a lot of climbers at a given event. If you’re down a climber, you need to make up 1.5-2.5 cycles over your opponent. Is that impossible to make up? Definitely not, but it’s a challenge, even distributed across all three alliance members. Also very notable, and a clever game mechanic by FIRST, if the opposing alliance has 2 climbs and I only have one… I’m only down 10 points if I can climb in the center and keep balanced. That’s not a death sentence by any means.

Tl;dr: I don’t believe Climb Roulette is back to the extent it was in 2017, but it is a very highly valued aspect of this game that every team should make a high priority.

27 Likes
2

Single climb centered is 40 points assuming you can balance the thing on your own. Much smaller L

2 Likes
3

Id say endgame in general has always been designed to make or break games. This year is no different.

3 Likes
4

You’ve got something I think you’re overlooking with this.

3 climbs is most likely 75 points.
2 climbs from a higher seeded alliance is likely 65 points.

Given the concern here mostly exists during playoffs, the alliance with 3 climbs is likely to be the 8th alliance captain plus two lower seeded teams that have a climb competing against the 1st seeded alliance (and perhaps 2v7).

In that scenario, I’d assume 1st captain and 1st pick have the ability to balance and are less likely to get a third member that climbs well. I’m also assuming the 8th captain and 8th and 9th picks are lower because, while they can climb, they’re slower as cycling and their climb isn’t as efficient as the higher seeded teams.

Given those assumptions, it’s unlikely the 8th seeded alliance is balancing. They’re losing the 15 points as a result and only taking the 10 point benefit. As that’s one realistic cycle, the change isn’t a vast difference here.

This only becomes a real difference if the lower seeded alliance is able to balance. At that point, the full climb difference takes effect. But, then you’re only looking at ~1.25 cycles extra for each of the top two teams. It’s a bit more. But, it’s not as bad as the game you’re mentioning where an extra rotor was likely the gain for the higher seed and that rotor was less points than the climb for significantly more work.

3 Likes
5

@JeffB you’re forgetting about parking. 2 climbs, level, and a parked 3rd robot is 70 points. Only 5 points less than 3 climbs.

6

I actually think that 3 climbs is likely to be balanced. Playing around with this simulator, if you put one full weight robot in the middle, there aren’t a lot of combinations of mass and weight that cause the coat hanger to not be balanced.

Side note: I expect to see very few triple hangs. I predict one alliance in playoffs (per event) will do it 50% of the time, and it’ll happen like 4 times per event in quals.

13 Likes
7

Is it possible the 8th alliance has the means to drive well enough to balance in the middle? Yes.

Do I expect that would take them more time if they go that route? A little bit.

Maybe I’m just a pessimist. I don’t see the 8th alliance being exceptionally talented with climbing together. Keep in mind, the 2017 example used for climbing was the year each robot had their own area to climb. They didn’t affect each other. Just climb on the side nearest your driver station so everyone could see their robot while trying to climb.

Will there be some out there to pull this off? I’m sure. Will it be the norm? I’m less convinced.

8

No, climb roulette is not back. A few points:

  • Climbing is only worth 20 marginal points. A climb which unbalances the switch is only worth 5 marginal points.

  • Climbing is significantly harder this year than in 2017 (I think this year is the hardest endgame challenge since I started FIRST in 2017, and 2017 was by far the easiest).

  • Climbing this year involves positioning and timing coordination among alliances bots in a relatively small space. Historically, tasks which involve cooperation between multiple bots are quite hard, and inconsistent.

  • Scoring was capped and sub-linear (pieces were less valuable as your scored more) in 2017. This year, the main scoring element is uncapped and linear. A climb is 2 cycles this year (depending on robot capability), whereas in 2017 it was usually about 5.

This wasn’t true in 2013, 2016, 2018, among recent games with an endgame. 2015 and 2014 didn’t have an endgame, as far as I know (I wasn’t around then).

This is a really good observation that may make triple climbing stronger than some teams expect. We definitely didn’t consider this.

9

I didn’t say a word about the likelihood of any specific alliance balancing. I think that balances in general will be relatively common because the bar is pretty forgiving, and I think that triple hangs will be relatively rare because they’re hard to coordinate.

10

Yes and no. The context implies the lower seeded alliances gain a benefit through the endgame. That’s what climb roulette was. You didn’t explicitly discuss the probability of any alliance.

If we’re looking at whether or not it’s back, we’re looking at the probability that the 8th seed alliance is likely to pull off a balanced triple climb while the 1st seed will only handle a balanced double climb. Each additional pairing of alliances makes this less and less likely (more picks to get a partner for the 7th alliance. less picks required for the 2nd versus the 1st).

11

I think the worse problem is in quals, in 2017, if you got matched with 2 robots that didn’t climb, you were pretty much bound to loose. I haven’t done the math this year yet but at first sight it shouldn’t be as bad

1 Like
12

Very rough estimates below:

  • 2020: Climb worth 1.5-2 cycles (+20 for climb, cycles worth 10-15)
  • 2019: Climb worth 4-6 cycles (+12 for climb, cycles worth 2-3) - note only 1 possible per alliance
  • 2017: Climb worth ~6 cycles (+50 for climb, additional 6 gears needed in quals to get the 4th rotor)
  • 2016: Climb worth 2 cycles (+10 for climb, cycles worth 5)
  • 2013: Climb worth 1-3 cycles (+10, +20, or +30 for climb, cycles worth 12) - note 30 point climb was hardest climbing task in FRC history
  • 2010: Climb worth 2 cycles (+2 for climb, cycles worth 1)

2020’s endgame isn’t that bad. 2017 is a clear outlier where one climb was worth a very strong gear robot’s entire teleop. 2020 is fairly comparable to every other year.

5 Likes
13

OP didn’t mention alliance seeding.

14

I would say that climbs are even more valuable this year than in 2017 because of the RP. You need at least one partner that is good at climbing, and if you don’t, you’re out of luck. This seems like a hybrid between climb roulette in 2017 and coopertition bridge balance roulette in 2012, where some robots spent half the match trying to get the balance and the 2 RP that came with it.

Last year had an endgame RP as well, but it was easier to achieve without much help from partners. A good team with a level 3 climber only had to rely on a partner to drive onto Level 1, which was a lot lower of a barrier. This year, a good climbing team has to play climb roulette with the match schedule and hope for pairings with other good climbing teams.

I’m pretty sure that most teams will build a climber and say it works, but there will be many other factors that determine whether they are able to do it well, such as driving ability, ease of lining up, and experience that can only come from practice. Sometimes, an alliance will have three teams that each claim that they can climb effectively, and if they view a triple climb as too risky, they’ll have to decide among themselves which two teams should try to climb and balance. With incomplete information, they could easily make the wrong choice. With complete information, they could still get unlucky and fail to balance, losing out on 40+ points and an RP. This sounds to me like true climb roulette.

15

Again, you’re also replying focusing on explicit things. If you’re looking at the idea of “climb roulette,” you’re looking at a situation that existed with the Steamworks game (and we’re exploring to see if it is likely to exist here).

There are two times when this showed itself:

  • Both alliances had 3 robots that could climb and at least 1 didn’t do so successfully
  • One alliance had 3 robots that could climb and the other had 2

The first didn’t tend to happen until you’re looking at the champs (district or worlds) level or a few very competitive regionals/districts. It’s not where the phrase was most commonly used.

The second has an implication that you’re either choosing to ignore by focusing on explicit mentions or you’re not understanding how it applies to alliance seeding. If it’s the former, shame on you. If it’s the latter, take a look at how bots are picked.

That season, the 50 points were crucial (especially given rotors were 40 if I remember correctly and were all or nothing meaning it took a lot of extra work to move 50 points ahead prior to end game). If there were 8 robots at an event that could climb, they’d almost certainly be paired in the top 4 alliances. If there were 16, they’d be paired in the top 8 alliances. If there were 20, the worst 4 climbers were going to end up being the 3rd robot on alliances 5-8 (starting with 8). Strategically, alliances couldn’t avoid picking these.

When you’re looking at whether or not this will exist in this season, the primary focus of the conversation is to look at how viable it is climbing will determine the outcome of a playoff match by itself. The most common scenario in which this is something to consider is the 1v8 (1 is the least likely alliance to have 3 climbers. 8 is the most likely to have 3 climbers. This is just based on the selection order and if climbing pushes the outcome, 8 will essentially HAVE to pick the 3rd climber).

I’m fully aware the OP didn’t explicitly state this. I’m fully aware the person replying didn’t explicitly state this. I’m also fully aware the most probable situation in which we’d see this is the 1v8 alliance seeding. If we want to see if it’s likely to occur, the first place to look is where it’s most likely to occur. After that, you can take it further to see how likely it is to spread to other pairings and see how likely you are to believe it’ll be prevalent.

I hope after this we’re done having a conversation about whether or not it was explicitly stated. It wasn’t. But, it’s still wholly relevant. If you’re not considering this, you’re not actively engaged in discussing what climb roulette was and whether or not it’s likely to occur this year.

Quick edit - Yes, there’s a qualifications aspect to this that affected ranking points towards getting to an alliance. This could have an impact on overall seeding changing who gets paired with who. Though, you’d still get to the eventual problem at the end where the scoring was weighted in such a way that it was often advantageous to be one of the lower alliances rather than a higher one. That’s a fundamental flaw in a game.

1 Like
16

Yes.

13 Likes
17

Good luck winning a match if you go up against an alliance that has multiple climbers and you only have one.

6 Likes
18

laughs in triple climb *

That’s the whole thought process behind our design, the climb roulette will destroy some teams.

*disclaimer : we might still fail miserably

3 Likes
19

Hell yeah climb roulette is back, and it’s somehow even worse with the balance aspect.

A third climb unbalanced adds only 10 points net over two balanced climbs. So not only is there the binary “did they make the climb or not”, there’s fierce debate over if the third climb is strategically worth it. Teams will be desperate to show their climbs in qual matches while other teams want them to not rock the boat to avoid costing a match or ranking point. It’s going to be a very intense and trying set of strategy discussions this year.

1 Like
20

IMO, in the vast majority of cases, the team that isn’t climbing for a balance shouldn’t even try to park… Yes, it’s more points if you do it right. But if you touch one of those climbing robots, you just cost 65 points and an RP.