Proposal: Judges to verify scrambles

I don't think a scanning app would be faster than the scramblers checking...because the app would have to take pictures or video of all sides, in a particular order, and would either need someone to operate or would need scramblers to place cube in there and wait for check.

I see what Dene talks about speed x accuracy. If someone can scramble a 3x3 with no mistakes but takes 1 minute to do so, we probably don't want him/her scrambling at a competition.

But yeah, scramblers need to check if they did it correctly. That's why we have the picture in the first place.
 
CubeStormers are, as said above, overkill, but using part of that idea, I think it might be plausible to make small, compact robots with motors capable of rotating cubes, connected to a phone to take pictures and check. Maybe that's still a bit far-fetched, just an idea.
 
Last edited:
10 seconds * 500 competitors = 1 hour 23 minutes
10 seconds * 100 competitors = 17 minutes

The time required for an extra layer of checking is not insignificant.
 
It's an extra 50 seconds per round. assuming each competitor has 5 solves, then that competitor will have an extra 50 seconds per round while the scramble is being checked.
 
I don't like adding responsibilities to judges. Currently, being a judge is a simple job that you can have anyone at a competition do, including parents and friends. While checking scrambles isn't hard, it is something that would take quite a bit of extra training, especially if they don't know much about cubing. While checking for colors isn't necessarily hard, it's just another barrier to letting anyone help out and judge, and I'd rather spend my time making sure scramblers are doing their job in the first place.
 
I rather like the idea of having pre-qualifications for scramblers. As stupid as it may sound, this could be loosely based on speed. I almost guarantee you a sub 13 solver scrambles correctly almost all of the time. If not that, you could just simply have a delegate verify the judge's scrambling ability (have them scramble a few cubes to prove that they can do it quickly and accurately). I'm in agreement of needing some kind of higher standard because I absolutely hate when I see people lose records due to the scrambler.
 
Why not have a separate designated scramble checker per round? Their job would be to act as a filter for all scramblers, and it would be really quick to do once you became familiar with the scramble after 2 or 3 inspections. It would also be much easier this way for it to be someone trusted by the judges, because its consolidated to one person
 
I don't think a scanning app would be faster than the scramblers checking...because the app would have to take pictures or video of all sides, in a particular order, and would either need someone to operate or would need scramblers to place cube in there and wait for check.

Two pics from opposite corners, doesn't matter which and in what order, should be enough.

One pic should be enough for almost all cases, too.

So the judge could lift the cover, snap a pic, put the cover back, and the app tells them "scramble 1", ..., "scramble 5" or "invalid", then the judge checks it's the correct number, and after the solve the judge snaps a pic of the timer and the competitor's QR code (instead of signature, and the judge doesn't have to sign, either, as they're logged in), and the result gets automatically entered into the database.
 
Last edited:
Image analysis, it's harder than you think.
Everyone has different cameras with different lightings. Yellows look like orange that looks like red.

I tried this to see if I could develop an 'auto reconstruction software' with my webcam. I soon gave up, the telly kept on flickering different light levels, it was enough to make me pack up my bags pretty quick.
 
To whoever said that speed doesn't play any role in scrambling, you can't be more wrong.

At every competition I've been to, there has been an underlying stress of maintaining optimum efficiency. We don't ask people who don't know the new Clock notation to scramble clock, because that would be inefficient. We don't ask people to scramble 3x3 if there are faster scramblers available. The goal is to keep track of the time and at least attempt to stay with the schedule.
Many times I've been handed 2x2s, done the scramble, and placed them right back in the cover without a second thought because I trust what my hands do, and usually we're pressed for time. The only solution I see to this is allotting more time for each event to allow for proper checking, or recruit more scramblers.
 
We do care about this problem. It's one of the most common problems, and especially bad for BLD.

If we just *say* that scramblers must check the cubes, it won't do very much.
While there are some nice ideas to encourage/require scramblers to check each scramble.

However, there are multiple things that could go wrong during a scramble:

- The scrambler does incorrect moves. (Depending on the move and where it was done, this could easily be caught by checking the picture).
- The scrambler does the incorrect scramble. (This happens frequently, and for this problem it's *not* enough just to tell the scrambler to check the scramble carefully.)
- The scrambler does a corner twist. (To guard against this for cubes, the scrambler needs to look at all 7 visible corners, *and* the hidden corner on the back.)

There are also fun details like the idiosyncrasies of different puzzles, and the sheer number of multi BLD cubes to handle.


I like obelisk477's idea of having a scramble checker (they could also verify whether puzzles are legal!), but most competitions don't have the resources for one. They'd also be very tempted to combine the scramble checker role with a scrambler role (or maybe have two scramblers and have each be the checker for the other), and we' have to be very clear about requirements.
 
To whoever said that speed doesn't play any role in scrambling, you can't be more wrong.

At every competition I've been to, there has been an underlying stress of maintaining optimum efficiency. We don't ask people who don't know the new Clock notation to scramble clock, because that would be inefficient. We don't ask people to scramble 3x3 if there are faster scramblers available. The goal is to keep track of the time and at least attempt to stay with the schedule.
Many times I've been handed 2x2s, done the scramble, and placed them right back in the cover without a second thought because I trust what my hands do, and usually we're pressed for time. The only solution I see to this is allotting more time for each event to allow for proper checking, or recruit more scramblers.

You have completely missed my point sir. This thread is called "Judges to varify scrambles" not "Scrambling is taking too long". I was dismissing things that are not effecting the issue which is being discussed.
 
You have completely missed my point sir. This thread is called "Judges to varify scrambles" not "Scrambling is taking too long". I was dismissing things that are not effecting the issue which is being discussed.

It's naive to look at this as a one-dimensional issue. A cubing competition is a dynamic environment, with multiple factors affecting multiple other factors all at once. If speed is no issue then my idea is definitely the best. If speed is a slight concern then Mr. Pochmann's suggested modification is actually a brilliant idea (although doesn't safeguard against everything).


Personally I like the idea of a "scramble checker". I might try implementing this at ausnats next weekend.
 
Two pics from opposite corners, doesn't matter which and in what order, should be enough.

One pic should be enough for almost all cases, too.

So the judge could lift the cover, snap a pic, put the cover back, and the app tells them "scramble 1", ..., "scramble 5" or "invalid", then the judge checks it's the correct number, and after the solve the judge snaps a pic of the timer and the competitor's QR code (instead of signature, and the judge doesn't have to sign, either, as they're logged in), and the result gets automatically entered into the database.

I really like this idea. It is just so logical, and *should* be relatively easy to implement.
 
I might try implementing this at ausnats next weekend.

Sounds like a plan.

We don't really have Judges walking back and forth at competitions. Here are some random thoughts on implementing the Verifier role with Runners:

Runner:
  • Go to the "IN" section of the scramble table;
  • Take a scorecard out of a cover and put it on the table;
  • Take the solved puzzle out of the cover and place it on the scorecard so that it's white/green or light/dark oriented etc;
  • Go to the "OUT section of the scramble table;
  • Pick up as many covers/boxes as possible and drop them off at competitor stations;
  • Pick up solved puzzles/covers on the way back to the scrambling area.

Scambler:
  • Pick up a puzzle from a scorecard (orient the puzzle correctly);
  • Look at the scorecard to figure out what solve the competitor is up to and which scramble to apply;
  • DON'T MOVE THE SCRAMBLE SHEET (OTHER PEOPLE ARE USING IT TOO)
  • Perform the scramble;
  • Verify that the scramble is correct;
  • Move the scorecard and the scrambled puzzle to the "VERIFY" section

Verifier:
  • Watch the "VERIFY" section for scrambled puzzles;
  • Check the scramble with your own scramble sheet;
  • Solve the puzzle and move it back to the "IN" section with the scorecard if the scramble was wrong; <-- or just interrupt the scrambler and get them to solve/fix the mistake (whichever is faster)
  • Put the puzzle in a cover/box and move it to the "OUT" section for a Runner to collect.

Tim.

EDIT: I don't think judges should check scrambles (beyond "is this scrambled?") with our current setup here Runners are used. Thoughts on scaling for large events?
 
Last edited:
Basically what TimMc has outlined above seems the only reasonable solution that works short and long term. One extra helper, vastly reducing error.

Nothing more frustrating than missing a PB average due to a bad scramble, when it was actually scrambled wrong.

TimMc's idea just needs decent scramble table space, and a quick briefing to runners, and it doesn't need extra technology. Whilst apps would be great, they would also introduce their own problems
 
You have completely missed my point sir.

Maybe because your point isn't worth anything.

@TimMc: Scramble checkers are an interesting idea. Given we have a fuller staff at the next competition BASC does, we can try it out. It doesn't slow thing down much because the output rate is more or less the same if accuracy is maintained. Of course, the only sacrifice is time to correct inaccuracies, but that should settle itself in time.
 
I definitely like the idea of a separate scramble verifier, if there is enough manpower. Of course the requirement to be a verifier should be much higher than the requirement to be a scrambler. It should be someone who has good attention to detail (perhaps a BLDer? :p) and who has been cubing long enough to have seen many different color schemes and to be decently fast at re-solving the cube if they need to.
 
Back
Top