What is going on with the “random” scheduling at the regionals? I’m here watching VCU, and the same teams are playing over and over and over…
116 vs 122, all 6 matches so far.
401 vs 405, all 6 matches so far.
343 vs 339, all 6 matches so far.
In NJ,
25 vs 11, 3 of 6 matches.
25 vs 41, 3 of 6 matches.
In PNW,
114 vs 192, all 4 matches so far.
488 vs 492, all 4 matches so far.
I think you get my point. If you dig further, you’ll find that the “random” software picked groups of adjacent team numbers, and just cycled them in the same pairings. This is NOT good for the competitions, as it just makes the strong strong, and the weak weak.
Could someone from the GDC PLEASE tell me that this will be changed for the following weeks of competitions? I feel very sorry for the teams competing this weekend…
148 also always has a rookie partner. Not only that, but every team with a 2 or 3 digit number always plays with 2 other 4 digit numbers. There also isn’t a single match with all 6 teams being 4 digit numbers, which should happen about 10% of the time.[/quote]
This is not a pretty situation. From my initial analysis, the algorithm seems to be grouping teams based on their team number, and cycling through these groups. As a result pairs of teams are seeing each other repeatedly. This is not fun for a team who has to play the same powerhouse repeatedly. Also, it gets very stale, very quickly. Hopefully the software is modular enough to handle a new algorithm, that can be implemented for the rest of the season.
This was definitely pointed out at the Winter Warzone scrimmage, and I was led to believe that it would be addressed before the regionals.
Our experience was to play all 3 rounds with the same 5 other robots. Robots that could only play defense so the scores were 0-0, 2-0, 0-0, no working ramps to try out between the six bots. This was Not the varied cross section you want when tuning up your robot at a scrimmage. And this could not lead to a fair shake for a good robot that ends up opposite the same power robots round after round.
At BAE, team 238 will play team 1153 8 times. Teams receive 8 matches.
At first, I thought our situation was bad (we played Buzz 3 times). I later found out that there are several instances of teams playing other teams (4+ times). We play either 1277 or 1247 in every single qualifier.
I never thought I would say this, but I want Hatch’s schedule algorithm back (and theres wasn’t particularly good).
On a positive note, the new field control can stay, although a timer would be more useful than a bar on the big screen.
PantherTech (292) is playing ThunderChickens (217) 5 times over the course of this weekend. Not to say that we don’t like playing you guys, but we’d just rather have you on the same side of the field. Yep, I don’t think this algorithm is up to pair and needs to be fixed. Hopefully people are playing attention and something can be quickly worked out.
Please Please Please. I am not trying to bash any team. My apologies go out in advance if anything I say can be taken in any negative way towards any team. The teams have absolutely nothing to do with the schedule and if asked they would all probably wish it were different. I followed the VCU regional today and after reading other posts in this thread I wondered if anything like that happened at VCU. I looked at the team in first place, I am not naming the team here because I am not trying to draw attention to them, just trying to point out that something is wrong with the algorithm that is being used. I know, I know, now everyone who reads this can go figure out who I am talking about. You have to be able to demonstrate proof of what you are saying and the only way to do that is to provide the information. They were against another team six of the seven matches they played That team is in 60th place. My guess is that both of thse teams would have much rather faced a variety of opponents instead of eachother quite so many times. Again, my apologies go out to each of these teams. In the four years I have been envoled with first I can not ever remember a time when the team I mentor would be with or against maybe a couple of the same teams for an entire regional.
These results are no different then the findings I had from Waterloo last year, mind you Waterloo is a very small regional, making it harder to mix up teams while still giving them a good enough break in between matches (was the reason given when I approached the issue). After approaching FIRST about this at a debriefing, I had the chance to see the code they used for Waterloo, it was in fact very well written. There was however a problem, this was that this program didn’t “reset” the algorithm after breaks such as the end of day or lunches causing teams to almost stay in a loop of teams they play with every match introducing one team into those they last played with as it ranked having a longer break for the team more important then who they were playing against. I changed this in that code adding a set number of resets, and saw a far more “random” selection between the days, and even better adding 2 more breaks. These two extra breaks are very accomplishable, if FIRST was to say add two 10 min breaks in the day one before and one after lunch, these breaks would also be logical as to allow the volunteers a much needed short break, allow any syncing problems with scoring to be worked out, and a multitude of other practical reasons. I relayed back my solution to the one that made the software, hoping these minor changes could be made.
It could very well be that this “looping” effect where you see teams play with or against others many times, is simply because they don’t have resets for the algorithm, or it could be because the algorithm was poorly designed, either way it is a problem FIRST knows about and I find it unfortunate for those teams who have been effected that they haven’t solved it yet.
edit: mind you there never was a problem with teams playing against a team all the time… say if they were stuck in the “loop” with said team, a few would be with a few against
The reason matches are not completely random is because of the algorithms used to “randomize” the matches. Although it may seem like random, adding little things like a minimum time span between a single team’s consecutive matches alters the randomness of the results. Now the results are no longer 100% random, but rather the sort of “equally-spaced-out-data” random that people generally associate with randomness.
It is not really completely random. You are more likely to play with the same people that you were competing with and against rather than someone else that is minutes out of your range due to the algorithm FIRST has in place.
this is absolutely true of the scheduling. the so called randomizing is making the strong stronger and keeping the weak at that same level of performance. This is not helping gathering any significant data during scouting eith:ahh: er. What are they going to do about this???
Well, it certainly makes scouting a little bit easier for every match.
401 plays against 405 EVERY SINGLE ROUND!!! Strangely enough, I’m not really noticing this when I drive though. Then again, we’re playing mostly defense right now.
