Picking up fuel from the ground vs. using hoppers/retrieval zone

[pmattin5459]

Originally Posted by euhlmann

Opposing teams who didn't design a floor pickup*
I assume even the most basic strategy discussion can come to the conclusion that hoppers can't be relied on

Yep. If our robot encounters any other shooters, we will dump every single hopper. It's a basic bit of defense that slows down many shooters' first pickup from a few seconds to however long it takes their floor pickup to fill their hopper (maybe around 5-10 sec?).

[g_sawchuk]

Originally Posted by pmattin5459

Yep. If our robot encounters any other shooters, we will dump every single hopper. It's a basic bit of defense that slows down many shooters' first pickup from a few seconds to however long it takes their floor pickup to fill their hopper (maybe around 5-10 sec?).

Sure, at a glance that's a fantastic strategy. However, let's look at some potential results.

If you do have a floor intake, I'll capitalize on the time you spend dumping the hoppers to squeeze in an extra gear cycle. That has the potential to make a big difference. Then, if you start scoring I will start cycling fuel in my retrieval zone, since you are feeding locations that I can load from. While you have the scoring lead, I have the gear lead (from when you were dumping hoppers).

Now, if you don't have a floor intake I'll start on gears, and just do gears continuously. I keep my one gear lead and you capitalize off of very little.

Not saying that hopper dumping (or better put, field reset's nightmare) is a bad strategy, just saying that it can be used against you.

[Siri]

I think folks are drastically overestimating how long it will take to finish a serious hopper race. In any worthwhile race, the only hopper that's "up for grabs" is centerfield (boiler side)--within a few feet and a few seconds of being on the path to a gear cycle. All the other hoppers shouldn't be a fight if the loader and defender are worth the effort--one each at least ought to be done by the end of auto. In fact the defender could decide to hit the centerfield in auto, perhaps leaving their own launchpad hoppers for their allies to hit on their way out (for a gear or what have you). The second offensive hopper should be hit by a loader's ally (upon good advice) no later than the start of teleop on their way to gear running and potentially doubling as the loader's blocker to the centerfield hopper. If the centerfield hopper is indeed still full by teleop, both parties are left to decide whether and how to race for it. But if you really can finish a significant portion of an entire gear cycle while the defender hits that button (which you have to pass approximately as well), it's probably an easy decision and you have bigger things to worry about. Like Einstein.

Separately, note that in a situation where a hopper dump tactic is worthwhile, relying on a single-gear lead countertactic necessitates both the ability to predict alliances' final gear totals and the luck that they'll end up on a rotor split. If that one gear does just barely let you finish another rotor, more power to you--but if you're so close to the edge, that usually settles the debate as to what you should be doing that match. It's also highly unlikely that the other alliance will end up one full gear cycle behind you in that case. To do so would be a mistake on their part that signifies they're really not on your level anyway--either they blew their time cushion or they should have stopped much earlier upon realizing they'll be a gear short. (The alternative case is your own alliance wondering why you ran a random extra gear that doesn't turn a rotor.)

From a strategic design perspective, a non-floor loader who poses a threat (i.e. made this as a competent strategic decision and properly executes) is more likely a fast gear runner who does fuel opportunistically from the retrieval station while grabbing a gear. They ought to be very good at the centerfield hopper race, though it's not a bad strategy try to beat them on your way past. Other than that and an autonomus load, they won't be preoccupied with hoppers in competitive matches (mostly because they really won't take long). Strategically that design decision screams high-cycle gear runner with some extra space/weight and design time. (If they're underperforming, of course, as with all defense the defender needs to reconsider pre-match whether it's worth bothering.)

[g_sawchuk]

Originally Posted by Siri

Separately, note that in a situation where a hopper dump tactic is worthwhile, relying on a single-gear lead countertactic necessitates both the ability to predict alliances' final gear totals and the luck that they'll end up on a rotor split. If that one gear does just barely let you finish another rotor, more power to you--but if you're so close to the edge, that usually settles the debate as to what you should be doing that match. It's also highly unlikely that the other alliance will end up one full gear cycle behind you in that case. To do so would be a mistake on their part that signifies they're really not on your level anyway--either they blew their time cushion or they should have stopped much earlier upon realizing they'll be a gear short. (The alternative case is your own alliance wondering why you ran a random extra gear that doesn't turn a rotor.)

I openly welcome you to risk 140 points during eliminations for the single gear that you don't complete because you wanted to dump the hoppers (or 50 for a hang that you might have not had time for as you raced to complete your 4th rotor).
I'm not quite sure as to why you think getting this one gear lead hurts you - you're working towards scoring while they're trying to counter you on a task you aren't targeting that match.
Please note that I never said if a floor intake or hopper intake was better - I think the others on this thread made that abundantly clear. Just providing insight with regard to countering hopper intakes.

We can apply plenty of levels of play to my suggestion - my thoughts are targeted towards a near-average level match up (perhaps slightly above).

[Chris is me]

Focus on your priorities. If a floor pickup takes time or resources or space away from gears and climbing, ditch it. If it doesn't, go ahead, that will have nonzero value. That's pretty much what it comes down to.

You'll be able to have some sort of ball game with human loading only. I don't think it's absolutely mandatory to floor pickup in this game.

[Siri]

Originally Posted by g_sawchuk

I openly welcome you to risk 140 points during eliminations for the single gear that you don't complete because you wanted to dump the hoppers (or 50 for a hang that you might have not had time for as you raced to complete your 4th rotor).
I'm not quite sure as to why you think getting this one gear lead hurts you - you're working towards scoring while they're trying to counter you on a task you aren't targeting that match.
Please note that I never said if a floor intake or hopper intake was better - I think the others on this thread made that abundantly clear. Just providing insight with regard to countering hopper intakes.

We can apply plenty of levels of play to my suggestion - my thoughts are targeted towards a near-average level match up (perhaps slightly above).

I expect that we mostly* agree with each other. My point was simply that people seem to be picturing one robot zooming around the edge of the field and hitting all the hoppers. A real race can't take nearly that long in reality, regardless of competition level. One robot getting to several dispersed hoppers in series at the start of teleop means by definition that either the other teams don't care enough to reach their nearby hoppers or they're broken/impaired robotically. Either way, even a defender that makes a habit of hitting every hopper they pass will reassess what they can do with their time. The point being that in matches where hopper dumping matters, the time trade-off will look much less like someone going far out of their way and more like situational defense in which someone hits me (or a hopper) while we're both working. Whether or not I choose to make that hit has a lot more to do with the specific field traffic at the time than any grand defensive or offensive philosophy.

The value of getting ahead of the opposition in gear cycling is a lot like other years, being chiefly determined by (A) your rotor predictions and (B) your time opportunity cost. Take a common competitive match situation in which we expect to be evenly matched with the opposition on rotors and climbing but may lag on shooting--it doesn't matter if it's Einstein or a regional qual. We can weigh this likelihood pre-match (with the final decision often based on autonomous scores); it's even arguably easier than in previous years given the massive time difference between finished rotors. So going a few seconds out of my way at the start of teleop in order to rapidly fill my hopper or prevent someone else from doing so may be a risk I decide is worth it. Perhaps I'll be wrong and have needed that time for a final gear that finishes a rotor. (This is why I said if you're close to your time cushion on finishing a rotor, the decision to gear is much clearer.) On the other hand, perhaps I decide not to hit the hopper, I still finish my last gear as expected, but then I can't get the last X balls that I need to beat their fuel score before the climbs. Or maybe I do win--or lose--comfortably, but I come up just a little under a 40kPa RP I really needed. In either latter case, hitting the hopper is a risk I should've taken and could've foreseen two minutes ago.

This is essentially like situational defense every year: sticking purely to offensive scoring at first--not detouring to hit a robot or a hopper--can feel like the conservative strategy, until it's late in the match and you realize you really needed to have made hit X while you were scoring in location Y thirty seconds ago in order to get/keep your winning margin. These situations go on and on every year, and they're among the most difficult jobs coaches have. The crux is that 2 seconds at time/location N do not have the same value as 2 seconds at time/location M.


*I will disagree on one point that, if one robot is in a position to want to go anywhere and dump (plural) hoppers in teleop of a what would be a competitive 4-rotor elim match, numerous people have already done something spectacularly wrong. A match like that (top-tier competitive and caring about hoppers) should have at maximum of one hopper any real distance away by the start of teleop, and that distance should be at maximum a few robot lengths off-course from someone who wants it. I may well decide to hit it if, as above, I expect to both need the 140 points and need that kind of fuel score. At an Einstein level nail-biter like that, trust me, you're not aiming for second place. It's very much go big or go home.

[GeeTwo]

Originally Posted by Siri

I expect that we mostly* agree with each other. My point was simply that people seem to be picturing one robot zooming around the edge of the field and hitting all the hoppers.

At least as 3946 is considering it, the point of the strategy isn't to deny the hopper-loading robot the first load (though that would be even better), but to deny or at least minimize the possibility of a second, third, fourth, and fifth "free" load. With a bit of drive practice or a cleverly arranged bumper/chassis configuration, a robot could spill all the hoppers on one side in a single strafe heading to the retrieval area, and the other side on the return trip. The bottom line is that if scoring (or even delivering) fuel is your primary activity, you need to be able to harvest them from the carpet or you'll find yourself moving on to a secondary activity early in the match.

[Siri]

Originally Posted by GeeTwo

At least as 3946 is considering it, the point of the strategy isn't to deny the hopper-loading robot the first load (though that would be even better), but to deny or at least minimize the possibility of a second, third, fourth, and fifth "free" load. With a bit of drive practice or a cleverly arranged bumper/chassis configuration, a robot could spill all the hoppers on one side in a single strafe heading to the retrieval area, and the other side on the return trip. The bottom line is that if scoring (or even delivering) fuel is your primary activity, you need to be able to harvest them from the carpet or you'll find yourself moving on to a secondary activity early in the match.

I'm not sure if you think we disagree. I don't think we do, but thank you for expanding the topic. I have three additions I'd like to make:

A) To clarify, my discussion in your quote is about a hopper "race" in which the issue is time tradeoffs versus opponent strategies. In this situation, the natural definition a "race" to me is limited to a two-sided competitive situation. Any alliance that cares to and can drive reasonably (and at higher levels probably load in auto) will reach their own closest hoppers before the opposition. It follows that any even hopper "race" by this definition can reasonably involve no more than one hopper, and thus racing closely resembles other field-positioning situational defensive choices rather than a loop. Races, which primarily start when someone wants to offensively use hoppers, msut end before multiple free loads by one robot. Essentially if you want hoppers (to score), they'll be gone. (So of course I agree if you're a primary fuel bot, you need to pick up and can't plan for more than 2 hoppers.) I also discuss other reasons to dump hoppers up front that aren't a flat out race, even if it means going out of your way (two seconds now isn't worth two seconds later).

B) In terms of whether or not you're going out of your way for a hopper loop, the field picture can be deceptive. Doing an along-the-wall run may mesh perfectly with your alliance's other auton-to-teleop (or later) priorities or it might not. For instance, this could commonly require a cross-over route with an ally that loaded from the hopper in auton. On many alliances who mesh well or otherwise aren't under too much pressure, cross-overs aren't really a big problem--for others they very much are. Separately, each leg of a hopper loop like this is a limited path on a constrained field in a limited time window (i.e. before the other guy decides they want a hopper) during which the driver is also prioritizing staying on the wall along with their other requirements. This has historically opened up exploitation opportunities for opponents of similar caliber as the driver.

C) To expand on the other point, if you find yourself not in a race when you thought you'd be, you do need to immediately ask yourself why. I use the word reassess--this doesn't mean don't do it, just consider what else you might have miscalculated in allocating your time and predicting the opposition. Did they screw up auton or the auto-teleop transition? Can you take advantage of it (something you almost certainly need to do right now before they regroup)? Are they better at gears than you thought? Do you need to rework your defensive timing/plan? etcetera.

C-1) Moreover, if you unexpectedly find multiple hoppers still available and want to dump them via a (still short) initial path you weren't mapped for, you need to communicate this properly. Failure to do so can and does cause far more traffic jams than it's worth. I've seen (and received) way too many robots T-boning their alliance partners when they try to unexpectedly blitz like this. Instead of the alliance zooming off in their respective lanes and reaping what should be "fast out of the gate" benefits from what can otherwise be a known teleop transition, multiple allies lose their initial field position and waste time bogging down. This year it could cause very unfortunately located traffic jams, especially if the opposition can exploit it. Pulling of a successful reroute benefits heavily from contingency planning, correct early autonomous predictions, and strong coach communication. You need to communicate correctly, and you need to be communicating with teams that will actually adapt to you--the latter especially is less common than you might think and is something you need to gauge realistically for that particular match.