Sebastián Pino Castillo Clock WRs Invalid

[Stefan]

Ok, I can't quite read your expression, but now that it's there, that's *my* fault

As a general guideline, I'd say nobody should post a probability in a serious issue like this one without specifying:
- for what exactly it is
- how it was calculated
Otherwise it's rather irresponsible.

Didn't know the OLL thing. Random fact or relevant here?

 

[qqwref]

Didn't know the OLL thing. Random fact or relevant here?

Not relevant.

What's important here is to notice that although a single solve with two adjacent solved clocks on a side (this is specifically what we should be looking at) is about 1/20, and thus likely to appear about every 4 official averages, having every single solve in an average be like this is extremely unlikely - on the order of having a solved 2x2x2 pop out of the WCA scrambler. And remember, this is on the level of an average, not a single solve and is an inherent property of scrambles, so there is only one opportunity per official round.


This is an important point to make as well: if, by a ridiculous stroke of luck, an entire average is so inherently easy that anyone who solves it will have a huge advantage over others, I'd expect the delegate to point this out and generate new scrambles - especially if that delegate is the only one at the competition who could potentially set a regional record. Not doing this is, at best, highly dishonest.

 

[Lucas Garron]

Ok, I can't quite read your expression, but now that it's there, that's *my* fault

I made some obfuscatory choices in condensing the code. Would have tried to be cleaner if yours wasn't already so clear.
(Basically, I generate all possible arrangements of 4 clocks, and perform pattern matching.)

As a general guideline, I'd say nobody should post a probability in a serious issue like this one without specifying:
- for what exactly it is
- how it was calculated
Otherwise it's rather irresponsible.

This sounds like a good policy. I'll try to hold myself to it: Sanity check the final result, and post clear code/math.

Could there be a standard way to specify things like this? (Actually, I have some good ideas, but they would take too much time and effort to implement compared to other important things.)

Didn't know the OLL thing. Random fact or relevant here?

Actually, this is what I first thought Carlos was talking about (due to the last clock). This is why I had a few calculations floating around in my notebook, and reported the wrong one.

 

[rubikaz]

Can you show us the calculation as well, please?

Ok. I do not know if you have observed that each clock have only one pin up and the 2 adjacent clocks aligned with the center are associated to this pin. So I have calculated it the probability of:

- each clock have one pin (and only one pin) up.
- each up pin has 2 associated adjacent clocks. These clocks have the same position than the center one.


The pins of one rubik clock have 16 positions. There are 8 positions with a pin up (4 for each face). So:

One clock have one pin and only one pin up is 1/2.
Each clock have one pin and only one pin up is (1/2)^5.

Suppose now that we get it. We have 1 fixed pin in each clock. The two adjacent clocks associated to this pin have to have the same position than the center one. Since 2 adjacent clocks have 12^2 positions and we have 5 rubik clocks, the final probability is:

((1/2)^5)/(12^2)^5=1/1981355655168

 

[Stefan]

if, by a ridiculous stroke of luck, an entire average is so inherently easy that anyone who solves it will have a huge advantage over others, I'd expect the delegate to point this out and generate new scrambles

When? I'd rather not have a competitor study the scrambles beforehand. And afterwards you'd have to repeat the round and that might understandably anger people.

(Basically, I generate all possible arrangements of 4 clocks, and perform pattern matching.)

Yeah yeah, I understood the basic idea, I just couldn't read all details. It's like French to me.

Could there be a standard way to specify things like this? (Actually, I have some good ideas, but they would take too much time and effort to implement compared to other important things.)

Don't know. I don't really care about the format, I just want people to tell *what* they calculated and *how*. As long as that's reasonably clear, I'm happy. But don't just spit out "1/1.981.355.655.168" without any explanation.

 

[Stefan]

I do not know if you have observed that each clock have only one pin up and the 2 adjacent clocks aligned with the center are associated to this pin.

Thanks for the clarification, describing your "what" and "how" very well now. But I think your basis is wrong: on the scramble sheet I've seen, the pins of scrambles 3 and 4 were not like that (the others are like you said).

 

[rubikaz]

Thanks for the clarification, describing your "what" and "how" very well now. But I think your basis is wrong: on the scramble sheet I've seen, the pins of scrambles 3 and 4 were not like that (the others are like you said).

Maybe my basis is wrong, I am not be sure. I obtained this information from Javier Tirado. Javier Tirado uploaded the following image in rubikaz.com (here http://www.rubikaz.com/foro/viewtopic.php?p=134785#p134785):

He says that this image represents the scrambles (i.e. the final position of each clock). The blue point of each clock is the pin that is up. I will ask him again about the image, maybe he uploaded an example.

 

[qqwref]

When? I'd rather not have a competitor study the scrambles beforehand. And afterwards you'd have to repeat the round and that might understandably anger people.

Once at least one person has competed and pointed out that the scrambles are far too easy, I'd expect the round to be repeated or nulled - by the same honesty principle that had Grzegorz Prusak agree to null one of his Square-1 singles that was misscrambled. It's more important to be fair and honest than to be efficient and inoffensive, and I think many of the existing rules and opinions reflect that.

Keep in mind, I'm just talking about entire averages that are so intrinsically lucky that it's far more likely someone planted them there than that they were generated by chance. So it'd be a good sanity check against the possibility of a cheating official, or a scrambler with good slight of hand skills. And you're right about nobody studying the scrambles beforehand - that's exactly why we need to consider this option, because if the scrambles were replaced before the round nobody could know. I don't expect such an easy average to ever legitimately happen, but the idea of replacing scrambles is important to consider.


Interesting hypothetical side question: suppose an honest delegate generates brings their own scrambles, and somewhere before one of the 3x3 rounds, at least one of the 3x3 scramble sheets is replaced by a scramble sheet brought by some corrupt scrambler (not a delegate or organizer) in a way that nobody else notices. Now, although the scramble is not blatantly easy, and in fact most people get a normal time on it, one of the competitors, in league with the scrambler, has pre-memorized an unconventional F2L order that gives them a LL skip. They then get a ridiculously fast time (let's say a low 5.xx). To all appearances this solve was just extremely lucky; video shows the scramble was correctly applied, they solve with a normal method and get a huge skip, and they act appropriately excited. Plus, nobody except the cheaters knows that the scrambles have been replaced. What should be done?

 

[Stefan]

He says that this image represents the scrambles (i.e. the final position of each clock). The blue point of each clock is the pin that is up. I will ask him again about the image, maybe he uploaded an example.

It's not just an example. It matches the scramble sheet I've seen, except the up pin of scrambles 3 and 4 is on the other side (that's the top-right and bottom-left scrambles on that image (weird layout)).

 

[Stefan]

by the same honesty principle that had Grzegorz Prusak agree to null one of his Square-1 singles that was misscrambled.

I think that wasn't because it was easy but because it was wrong, which I consider something entirely different.

 

[qqwref]

When I said "by the same honesty principle", I meant that it was removed because it would be unfair to keep it, and thus similar situations should be dealt with similarly. We only replace/DNF wrong scrambles when they're unfairly easy, which suggests that it's not because the situation involves a wrong scramble, but because the situation is unfair to others.

 

[Stefan]

Yeah I understand, I just disagree that naturally occurring easy scrambles should be invalidated (unless it's in the regulations, but I think it isn't).

Edit: And it shouldn't be a delegate's task to make up rules for what is "too easy" on the spot. If we do this at all, it should be the same everywhere, otherwise *that* would be unfair. Since the WCA provides the scramble programs already, any such rules could be built into them. In my opinion, all scrambles allowed by the WCA regulations and created by the WCA scramble programs are ok, and must *not* be invalidated.

 

[irontwig]

And if a delegate is able to remove too easy scrambles should (s)he also be able to remove too hard scramble too?

 

[qqwref]

In my opinion, all scrambles allowed by the WCA regulations and created by the WCA scramble programs are ok, and must *not* be invalidated.

So then there is nothing wrong with Sebastian's scrambles...? After all, they're valid scrambles - the WCA scramble program could have created them. We can't know for sure. Even having the same scramble as your WR single again a few comps later is possible by luck alone. (The point of removing too-easy scrambles is NOT to work around the official scrambler - it's to help deal with possible unprovable cheating situations.)

And if a delegate is able to remove too easy scrambles should (s)he also be able to remove too hard scramble too?

I don't think that unfairly difficult scrambles exist. We know from the distributions of optimally solved twisty puzzles that they tend to share a long tail in the direction of fewer moves, and a much shorter tail in the direction of more moves. The same thing happens with human methods: each step can be skipped, which saves a lot of time, but there are far more cases where it isn't skipped, and those cases are all similar in difficulty.

 

[rubikaz]

It's not just an example. It matches the scramble sheet I've seen, except the up pin of scrambles 3 and 4 is on the other side (that's the top-right and bottom-left scrambles on that image (weird layout)).

Hi, I can confirm now that you are right. But it seems that scrambles 3 and 4 were prepared but there was a mistake with that scrambles: the position should be the correct one (for the skip) if you interchanges the front face and the back face. Of course, I can not be sure about it but I also have read that Sebastian told to the judge (twice) that the scramble of the clock he had to solve was not the correct one. They checked it and it was correct. So I think that maybe he told it with scramble 3 and 4 because he though that the position of the spins had to be the good one for getting the skip. Of course, I can not affirm it. I will ask about it. Edit: the blue text speaks about other competition. Sebastian did not get any WR there.

Anycase, I have calculated again the probability of getting 2 adjacent clocks in the same position that the center one. If I have not make a mistake, it is:

- 1 clock: 21658847 against 429981696 (approx. 1 against 20).
- 5 clocks: 21658847^5 against 429981696^5 (aprox. 1 against 3 millions).

 

[rubikaz]

but I also have read that Sebastian told to the judge (twice) that the scramble of the clock he had to solve was not the correct one. They checked it and it was correct.

I have been looking for what I have read. Ok, I do not know it it was twice or once. It was written by Alvaro (a Chilean cuber) here: http://www.facebook.com/note.php?created&&note_id=171073149620824

EDIT: This text speaks about other competition. Sebastian did not get a WR there.

-En la participación de Bahía Inglesa ocurré un hecho trascendental, que tiene como testigos a dos participantes mas que me han pedido mantener oculta su identidad (cosa que no estoy de acuerdo), y a mi. Estando en la resolución de clock, mientras Sebastian inspeccionaba dentro de los 15 segundos, empieza a retar a la Gabriela que era la Scrambler, diciéndole "está mal scrambleado". Este hecho llamó inmediatamente mi atención, porque.. ¿el tendría como saber si estaba mal scrambleado?.

He says that during the 15 seconds for inspecting the rubiks clock, Sebastian told to the person that scrambled it that the scramble was not correct.

 

[Stefan]

So then there is nothing wrong with Sebastian's scrambles...?

Right, I don't see anything wrong with those scrambles themselves. If they were created by the WCA program in the normal way (*), I think they're totally fine and you're not "highly dishonest" if you use and keep them.

(*) Of course you must not produce millions of scrambles and pick easy ones, you must use the first you get and not analyze them.

The point of removing too-easy scrambles is NOT to work around the official scrambler - it's to help deal with possible unprovable cheating situations.

That's an interesting point, hadn't thought about it from that perspective. Not sure how good that is, though, it only prevents rather naive cheating and might give a false sense of security. But again, I'm not totally against removing too-easy scrambles, I'm just against doing so in a highly subjective way, without shared regulations for it. And against calling someone "highly dishonest" for using what WCA provided.

I don't think that unfairly difficult scrambles exist.

Let's give Feliks scrambles where color-neutral cross needs 9qtm and see how well he can solve cross and look ahead to F2L

He says that during the 15 seconds for inspecting the rubiks clock, Sebastian told to the person that scrambled it that the scramble was not correct.

That alone sounds really bad (no matter what the scramble was), as he shouldn't be able to tell beforehand whether it was scrambled correctly.

 

[hic2482w]

He says that during the 15 seconds for inspecting the rubiks clock, Sebastian told to the person that scrambled it that the scramble was not correct.

"Hey! This wasn't the scramble I ordered!"

 

[Vincents]

I have a slight problem with "easy scrambles". Doesn't it depend on what method you use? A scramble may be incredibly easy for a Roux-er but incredibly difficult for someone using MGLS. How do you avoid discriminating? You could protest that maybe a scramble that is 10 moves from solved (for 3x3) is "easy", but where do you draw the line?

 

[Sa967St]

I have a slight problem with "easy scrambles".

Yes, you do.