Rank or EPA to win an event?

Competition Rules/Strategy
21

Please please please heed this advice. EPA is only as good as its input data, which is extremely noisy and coupled to a lot of external factors. If you want to compare teams accurately there is no substitute for scouting the matches manually and tallying individual robot performances yourself.

10 Likes
22

Great question. Always focus on skills to participate in both rank points. Then focus on match plans that give you rank points with your qualification alliance. You can’t win outside playoffs, so rank to control your fate. As stated above, autonomous and endgame skills are always a must as well.

Now it gets tricky on selecting a design strategy, especially against the teams at your events. For example, the last two had a way to particate in rank points by scoring high volumes of game pieces low. You may be able to rank that way, but may not win the event. So you would have to get to champs on a wild card.

These design decisions need to be made within your team resources. But consider the design journey of 4481 in 2022. They were a super fast low goal bot at their first competition and worked up to high goal through the season. Bottom line, you may want to consider fast and low to differentiate yourselves.

David

3 Likes
23

IMO, a team’s priorities should be:

  1. Don’t end up a “DNP”.
    1.1 Don’t be obnoxious/overbearing/etc
    1.2 Be honest with alliance mates about your robot’s capabilities
    1.3 Don’t lie (eg if you say you will play defense, do so)
    1.4 Don’t get any Yellow/Red cards or lots of Fouls
    1.5 Don’t be unreliable, esp. in embarrassing ways (eg battery falls out)
  2. Winning matches
    2.1 know your role in the alliance and play it at your best
    2.2 know that 2 RP for a win >> 2 RP in a loss
  3. Getting additional RP
    3.1 Drive coach always knows your RP status
    3.2 Be ready to modify play to get RP
    3.2.1 example: you have a good points lead and have enough links for that RP, but you need all 3 to dock/engage due to auton fail – maybe start climb a little bit earlier to insure success
    3.2.2 example: you only need 2 to dock/engage for RP, but are short for the links RP – don’t attempt triple climb, one robot keeps scoring cubes/cones all during endgame
  4. Look amazing
    4.1 “run up” the score after locking in all 4 RP
    4.2 have multiple reliable auton routines
    4.3 bring a boss-looking robot
27 Likes
24

I 100% understand this but I still think I have to go into the event with the goal of winning. Finalist wildcard is just a slight failure of that goal and is on the direct path to it. I like your declining strat, that is something I’ve thought about and might have to happen if the real goal is worlds. And you are right about the whole EPA thing, I should have used “scoring” instead. EPA is just a good base way to get an idea for scoring power. Does the FRC community have a word or acronym that represents pure scoring power?

3 Likes
25

I think typically people call it OPR (Offensive Power Ranking/Rating) you can see it on the blue alliance. But the issue arises that people may confuse you using the term as the stat TBA is reporting vs the actual ability of the robot.

1 Like
26

I don’t think this is a useful way to frame scouting. There are ways to attempt to calculate overall scoring metrics but they’re fraught, because the problem is severely underdetermined from the match data alone.

Instead of focusing on metrics, focus on on-field performance. Don’t confuse the symbol for the substance.

1 Like
27

A bit of a semantic point, but it will be power rating, ranking would imply an ordinal rank in most scenarios, which is not the case here.

You are correct that the terms are used somewhat interchangeably when folks are not using the acronym. As all a lot of teams care about at the end of the day is the order when sorted by OPR (thus rank), and less about by how much distance between each team position.

Carry on as you were :slight_smile:

1 Like
28

This entire thread just serves to remind me about how lucky we are to be in a District, and to hope that more FIRST regions make that switch soon.

7 Likes
29

It can be a huge benefit and it can also backfire to choose pure points output(cube output in this game). We(842) gave up rank 1 in 2018 by choosing not to have a buddy climb and focus everything into scoring on the scale. And whether it be finishing our auto modes or finishing the robot, we were always barely on time for our next event so if we had to add a more complicated climber, our robot’s offense would definitely have been effected. In regionals this worked out very well for us, but it really fell apart at champs and not being in control of our destiny really hurt us. I do believe we would have had different results in another division so in a way, going for pure scoring is riskier but potentially higher reward. But it depends on the game too whether pure scoring and RP is aligned and by how much. It also might be less risky now than in 2018 because the overall quality of robots has skyrocketed.

3 Likes
30

Looks like y’all are in mid tier hell. This is hardest portion of a team’s development arc, and requires a long term outlook to get through.

@john3928 list is correct for what required at an event but as you well know your performance ceiling is decided long before the event.

To that end I would suggest the primary difference between a consistently high ranking team and one that is stuck in mid tier hell is the team’s “Iteration Velocity.”

How quickly can your team come up with an idea, test it and implement it on the robot. Getting super fast at Iterating your teams robot will enable your team course correct as needed during the seasons such that you are not making the same mistake event/event. A really fast Iteration cycle also ensures your robot is done with enough time for solid driver practice and advanced software dev.

Solid driver practice is also critical before your first event to identify problems with the design and ensure they get fixed before the issue caused your team to lose a match.

The number 1 thing is to never make the same mistake twice.

11 Likes
31

I agree. And the goal of these iterations should be (1) increasing the reliability of your scoring cycle, then (2) increasing the speed of your scoring cycle.

I am assuming as a “mid tier” team you are already usually going to be designing for the “high goal” (whatever form that might take).

If you are a lower tier team, step (0) is an honest assessment of what scoring tier is within your capabilities. In (almost?) every game, an awesome, reliable, fast low-goal scorer will be valued far above a janky, unreliable and/or slow high-goal scorer. Nonetheless once you are accomplished at that, it’s very normal to want to start building high goal scorers, just be aware you may experience some “performance reversion” as you start up that learning curve.

6 Likes
32

One thing that slipped my mind is that in the regional model, the second pick does not automagically qualify for Worlds. Your chances do improve at later events, though.

33

Hi Drew,

I took at look at 1388 both events this year. Looking at the teams in both events it is a pretty low chance that you would be able to be ranked 1 or 2 at the end of Qualification matches. You have 4 very high tier programs at Ventura and 2 again at Las Vegas. Looking at most years the 1st Alliance usually has about at least a 50 percent chance of winning the event. The 2nd has about a 20-25 percent and then it goes down from there. Since there is such tight competition at the top end for your events it makes sense for 2 routes. This is assuming the goal is to get to Worlds. The best idea would to be to get on the First or Second Alliance. For Ventura, it is unlikely that you can rank high enough to be in the 1-3 spot. Most Likely you will be 5-19, which is showing you will most likely be picked in Elims. The best chance for this event would be to go hard on pure scoring and have the top teams fight for the ranking points. This will give your team much more practice in scoring like an Elimination game and give you the highest chance to be the 1st or 2nd overall pick. If a Scorched Earth happens at this event your teams odds go up significantly. For Vegas, I would say since the field looks different then there is a chance that your team can rank 3. This would enable you to be the 2nd alliance captain. At this point your odds are still slims, but since you are a captain you can pick a team thats only goal is to beat the 1 alliance. Look at defense or other ways to make the game messy so they cant play how they want to and design a unique alliance with the goal of taking them down. If the 1 alliance already has a slot from worlds and you are the captain then also going for wildcards here can be smart and draft a team that is built to get second. I think this is a great question that alot of teams assume the RP is the right answer but there is alot to consider depending on the team.

1 Like
34

Our team had a conversation with someone from 862 in 2022. What it boiled down to was that our team would always aim to be the 1st seed, and they would try and be that second pick. They would specialize at one or two things while we would try and be able to do everything. I do however agree that its a risky bet.

Its worth noting that they have more banners, but their successful seasons correspond nicely with ours, so it seems both strategies work.

So I would say EPA first, because that helps you stand out to become a pick, and then rank. You can win without a good rank or EPA, but unless you are in the top 10 every event, rank doesn’t matter as much.

35

It’s very important to set healthy goals for your team and yourself. When you are an underdog, it’s critical that you know who you’re playing against, and what they’re good at. In college, I mentored 5254 for a few years - we had 12 kids if everyone showed up, 4 mentors (3 of which lived > 45 minutes away), a drill press, a chop saw, and a constrained practice space. Is this the type of team that you might expect to win events? I don’t, but I like to think we came pretty close anyway, and I’m very proud of it. We certainly learned how to fight as an underdog.

Looking at 1388s 2024 events, your events are difficult but certainly not impossible. You have a reasonable schedule (wks 2/5) and you’re against some perennially good teams, but not so many of them that you should expect to be boxed out of finals no matter what.

To be unfortunately blunt: unless your team has completely overhauled yourselves over the offseason, I do not recommend building a 4 RP robot in 2024. Period.

Luckily, even if you want to win, you don’t need to build a 4 RP robot.

Obviously, every game is different. RPs can’t necessarily be compared year to year. Some games are easier to get 4RP (2018, 2023) than others (2017). Some games are middle of the road (2022, 2019).

Consider the relative difficulty of a 3RP robot that scores a lot of points (compared to a 4RP robot) in any given year. If you’re having trouble remembering, try:

You will notice that all of these robots did quite well (as an understatement). All of these teams spent a significant amount of effort on drive practice - 6329 was practicing this year before we had even assembled our chassis.

Drive practice is the single most important aspect of winning.

Build simple. Be scoring 4-5 weeks after kickoff. Practice more than you think you need to. If you go in to an event, win 80% of your matches, get a single bonus RP in every match, you will gain 26 / 40 RP for a ranking score of 2.6.

At every single event 1388 has attended since 2016, when RPs were introduced, that would put you in the top 8. At 9 of 13 of them, it would put you in the top 4. And you didn’t even build a robot that was “supposed” to seed!

You don’t need to win your first year. Set challenging - but realistic - goals.

14 Likes
36

What do you mean by scorching? I haven’t heard that used this way.

37

Frank does a pretty good job explaining the term in this blog post:

4 Likes
38

Strongly disagree with 2018 being easier to get 4rp and 2019 only being middle of the road. 2018 4rp was essentially “did you have a buddy climb?” which made it easy for the top level teams who did have one, but almost impossible for teams who did not have a buddy climb. 2019 was also incredibly hard in terms of RP difficulty not only because the rocket RP was 12 cycles, but because was extremely easy for the opponent to deny. Effective defense in 2019 was super easy to play because it only took sitting on of the rocket to block off half the hatch panels and pushing the opponent two inches to deny a scoring attempt. Even an inexperienced defender could play effective-enough defense to deny the rocket RP.

1 Like
39

It depends on the game.

This past year if you were a team that focused just on low goal scoring (drop one in Auto and run ~20 second cycles in tele-op) plus do an auto balance at most of the regional / district event level events I’ve seen you’re probably going to walk away with 1 RP every match for the balance, get 3 of the (usually) 4 links you need for a second RP. A lot of times getting the auto balance, 15 link points, and 18 points (or whatever the game piece score was), and balancing again at the end of the match would probably make you you one of the three highest scoring robots on the field during qualifying depending on how competitive the event is.

We consider ourselves a higher resource team and got way behind on our robot design this year and ended up focusing heavily on low goal scoring during/after our week 3 event because mathematically we weren’t leaving too many points on the table doing focusing on low goals but we were killing ourselves in ranking points trying to fill the high row plus a mid/low. So, we weren’t clever enough to both do the math and buy-in early in the season but teams who did figure it out and started low this season, got your auto balance working, and then maybe built an arm to do something on mid-level later in the season you’d have been an effective robot. This year’s game was about game pieces movement by the time you got to late season events / district champs / world champs. If you could cycle 6 pieces and fill up the low level this year you’d have been valuable to both win the match in playoffs and get a high ranking position.

Now compare to 2019 where there were robots that could only do a high climb that would rank very high because of the RP rules. That year there just weren’t enough points on the high climb to make up for not doing anything else --so you could rank high and the best bet you had is hoping you’re high enough to scorch the field or else you were going to get knocked out early during the playoffs.

Now, the hard part of all this taking a deep breath the day or two after kickoff, understanding the game early in the season, and making good assumptions on cycle times and stuff like that. If you’re a low resource / “average” / "mid-tier team taking the time on game analysis and you get alignment with your teammates and mentors to limit the focus on how you think your team can best help an overall alliance vs. worrying about “what if we don’t have a partner who can do_____” is most of the battle we ran into before we became a team that could everything (fairly) well. I’d argue even if you’re “wrong” when it comes time to compete being really good at something is better than having a robot that doesn’t do anything well. Like if your one of the teams in 2019 that decided to only climb those teams still at least put themselves in position to be picked / get in the playoffs and have fun cheering on your team for at least a few playoff matches.

I think Team 4481 has done a good job over the past few years with their game analysis that gets posted in their Open Alliance thread and even reaching out to decent teams in your region for advice might be helpful. I bet by Tuesday / Wednesday of kick-off week I’ll start getting text messages from some of the mentors on teams in the area asking what my / my Team’s opinion is if I was in their in shoes (knowing what their resources are, historical issues, etc…).

6 Likes
40

I really don’t agree at all. “Tarzan climbers” (where you attached a rope, then drove out of the way before winching) became extremely potent, even by week 3. On my team at the time, we achieved a 3.3 RS in a local event (thanks to several local teams having these climbs), and a 3.1 at champs*, after intentionally ditching our buddy climber because tarzan climbers were so potent and available. In fact, 1241 nearly seeded first in our division beating out several of the best buddy climb robots in the world (2590, 319) and tying RP with another (2056).

I don’t agree it was easy. It was easy if you let them. But a few teams found ways to not let it happen as often as people like to think.

If you see a defender parked in front of, at most, 30% of the field, and you attempt to score there despite the other 70% being guaranteed undefended, then maybe it’s a tiny bit your fault (the royal you, not Alex_Y in particular).

3 Likes