I just learned that apparently there have been cases where scrambles were discarded: http://www.speedsolving.com/forum/showp ... stcount=50 http://www.speedsolving.com/forum/showp ... stcount=51 I was under the impression that we allow all scrambles, not discarding any. [b:2rmsj0ke]I would like the discarding of scrambles be standardized (which could also mean forbidding discarding).[/b:2rmsj0ke] Right now I'm not sure whether/how I'd like it, but there should be clear and easy to follow rules for this, not individuals discarding scrambles because they "feel" they're not good, without reasoning agreed upon in advance by the community/WCA.

I don't really like the idea of discarding scrambles. However, we may encounter some scrambles that have a so short solution that everyone can find it in a second. For example, I would not like to see in a competition a 2-move solution for the 3x3, it would kill the single WR. So, how could we find a limit that will be accepted by most of the people ? We could also imagine scrambles that are easy not in the number of moves of the solution. Imagine a 5x5 scramble with all centers solved ? This is a much more complicated situation, we could not define a theoretical limit for what we should discard and what we should not. So I would let the main judge or the wca delegate decide during the competition, before the event has started.

[quote="Clement Gallet":1vuti6d0]So I would let the main judge or the wca delegate decide during the competition, before the event has started.[/quote:1vuti6d0] Decide on what? On rules for discarding scrambles? That's clearly the job of the WCA and to be done in advance and in public (i.e., right here in this forum and then published in the official regulations). Or on specific scrambles? If there are rules for the decision, these should simply be followed, and if there aren't any, then those guys would make arbitrary decisions quite possibly not followed in other competitions. Inconsistent and thus bad. I think I'd vote for either no discardings or only trivial ones, let's say "scrambles solvable in three moves or less", which even non-cubers could solve.

[quote="StefanPochmann":3r6ikyjv][quote="Clement Gallet":3r6ikyjv]So I would let the main judge or the wca delegate decide during the competition, before the event has started.[/quote:3r6ikyjv] Decide on what? [...] on specific scrambles? If there are rules for the decision, these should simply be followed, and if there aren't any, then those guys would make arbitrary decisions quite possibly not followed in other competitions. Inconsistent and thus bad.[/quote:3r6ikyjv] Yes, on specific scrambles. But I realize it would be a source of possible problems because of the inconsistency you are referring. Is there any good way we could avoid those easy scrambles ?

Would you want to get back to discussions on for example the TP forum about banning 'lucky' scrambles? It's hard to define easy at all. Where do you draw the line and on which criteria do you base your line? IE a 2x2 scramble of 5 move possible solution Derrick and me once had while racing 5 move scrambles was extremely hard, while we had some really easy ones with 8 move scrambling. So before we start discussing this, how do we actually want to decide on this subject? If we don't guide this I forsee a big discussion thread with info that's already discussed again and no real solution.

[quote="Erik":qaiaj53a]So before we start discussing this, how do we actually want to decide on this subject? If we don't guide this I forsee a big discussion thread with info that's already discussed again and no real solution.[/quote:qaiaj53a] I agree on Stefan's idea about discarding up to x moves solutions. we could set x := n for the nxnxn puzzle. For other 'lucky' solves, as I don't have any idea how you could define them, I would not discard them. So if anyone has a *good* proposal about discarding them (if it exists), I'm ready to heard. If not, just don't say anything. And Erik, this is still the wca forum, so even if we had a discussion about that somewhere else, we still have to talk about it here, as it's the right place to do so.

Good point, Erik, that TP discussion was really useless and I don't understand why people kept talking about it for so long. Although at least until now this discussion has been on a different level, I'm mostly concerned about consistency. I might be quite pissed off if I'm denied a three moves scramble in one competition while someone in another competition is allowed to get one, just because the organizers feel differently about the scrambles. You allso reminded me of [url=http://www.twistypuzzles.com/forum/viewtopic.php?f=8&t=10906:3nnarsf3]this poll[/url:3nnarsf3], a reaction to said discussion after your 7.08 WR. It asked [i:3nnarsf3]"Satisfied or dissatisfied with the offficial regulations concerning lucky solves and scrambling?"[/i:3nnarsf3]. So far it's 27-0 for "satisfied" against "unsatisfied". Since the regulations as far as I know don't talk about lucky solves or discarding scrambles, I think this means nobody (at least of those 27 who cared to vote) wants to discard any scrambles. And I think under the current regulations, the abovementioned discarding of scrambles was wrong.

[quote="Clement Gallet":18v41x04]I agree on Stefan's idea about discarding up to x moves solutions. we could set x := n for the nxnxn puzzle.[/quote:18v41x04] Ugh, solvability in seven moves of a 7x7x7 can be quite hard to determine. That doesn't fit my "trivial scrambles" idea. If we do this at all, it should be easy to do so everybody can actually follow the rule. But new suggestion: [b:18v41x04]Discard scrambles that would count as solved or solved with penalty.[/b:18v41x04] This way we'd not introduce a new decision process but reuse an already used/familiar one. And it would prevent potential trouble in the case where someone accidentally gets an unscrambled cube from the judge (happened to me before, both as competitor and judge) and insists on keeping that scramble arguing it's perfectly ok.

[quote="StefanPochmann":2fqm5kyv]But new suggestion: [b:2fqm5kyv]Discard scrambles that would count as solved or solved with penalty.[/b:2fqm5kyv][/quote:2fqm5kyv] Oops, I forgot something. Let me correct the suggestion: "Discard scrambles that would count as solved or solved with penalty, [b:2fqm5kyv]but no others[/b:2fqm5kyv]." This aspect is important and applies to other possible discarding criteria, too. A rule saying [i:2fqm5kyv]"Discard if X"[/i:2fqm5kyv] isn't enough, [i:2fqm5kyv]"else don't discard"[/i:2fqm5kyv] is necessary as well. Otherwise the inconsistency problem isn't solved at all. Some organizers might discard only if X, others might discard more cases.

[quote="Clement Gallet":2qwsc5ja]For example, I would not like to see in a competition a 2-move solution for the 3x3, it would kill the single WR.[/quote:2qwsc5ja] This will newer happen because only 24*18 scrambles out of 24 * 18^24 are 2 moves. OK, we use random setup nowdays and maybe my figures are wrong but you get the picture (it was far much harder to calculate figures for setup probability to become 2 turns, maybe something for you Stefan? =)

[quote="StefanPochmann":67nbtu2m] "Discard scrambles that would count as solved or solved with penalty, [b:67nbtu2m]but no others[/b:67nbtu2m]." [/quote:67nbtu2m] Agreed. Like Erik's quote on speedsolving, getting lucky is not a crime.... Averages are more indicitave anyways.

I agree with Stephan on the point that it must be clearly defined by WCA. One scramble can not be accepted in a competition and discarded in another one, just because the main judge or the WCA deleguate is different. My personal opinion about this is to never discard one scramble. As the limit is too hard to find for me, and to have a right sense. I don't care about killed singles because a single has no sense and only matter for journalists.

[quote="Kenneth Gustavsson":2tum2jms][quote="Clement Gallet":2tum2jms]For example, I would not like to see in a competition a 2-move solution for the 3x3, it would kill the single WR.[/quote:2tum2jms] This will newer happen because only 24*18 scrambles out of 24 * 18^24 are 2 moves.[/quote:2tum2jms] Yes, and we will never see a huge meteorite hitting the earth and causing many species to disappear.

In 3x3x3 scrambling, the probability of having a 'ridiculously easy scramble' is so low, that we haven't begun to consider what we would do if we come across a scramble that gives the first two layers solved. However, competitions are becoming more common, and the cubing population is increasing, whereas perhaps this is not a problem in 3x3x3, we do have Clement's proverbial meteorite in the 2x2x2. With larger cubes, it's very difficult to determine solveability, and really, the larger the cube gets, the less it matters, right? We know that this matters for 2x2x2, and it might matter for 3x3x3, but it's influence quickly diminishes almost exponentially. (not used an a mathematical sense, I don't know how the probability curve looks) Who would determine that the 2x2x2 scramble can be solved in 3 moves? I guess it is unlikely that competitors notice it if the organizers don't... but don't you see a situation that's possible where the scramblers aren't really good cubers and can over-look such a scramble? What about 4 moves on a 2x2x2? Something like R U R' U'... which I think I've seen in competition before.

Really, no scrambles should be discarded. If you're worried about something being "easy", then it should be blocked by the scrambling program. Anything that comes out of that program is legal. The reason I say this is that we take away the human process. I can look at a 2x2x2 scramble, but I wouldn't be able to tell you if it's a single Guidmond-step away from being solved.

Actually, that's a brilliant idea. If you're going to discard anything, it should be in the scrambling program. I'm not necessarily saying that things should be discarded or not, but if it is implemented, it would be nice to reduce the human element. Isn't that why we said not to discard solves in the first place? A solve easy for a Fridrich solver is not the same for Lars Petrus. I think at WC 2003, the organizer asked people for the color of their cross, so that the scramble would make that first step more difficult. Of course, this is unfair as it penalizes only a specific subset of the cubing competitor population.

I would not mind to filter 2x2 and 3x3 scrambles. But only the ones with an extremely low number of direct solving distance. For 2x2 my guess would be 4 or lower. For 3x3 my guess would be 9 or lower. But I would not mind to not change anything at all. Ron

I think that a best of N round where one scramble has a 2-move solution is ridiculous. I think this situation needs to be avoided by discarding the scramble with a 2-move solution (or whatever cutoff point is deemed best) or having the scramble program automatically discard such scrambles. Either that or best of N format should not be allowed on puzzles where such easy scrambles have a significant chance of occurring.

[quote="Ron":3hwowfw7]For 2x2 my guess would be 4 or lower. For 3x3 my guess would be 9 or lower. [/quote:3hwowfw7] Even for these though, there are some "easy" scrambles... so I don't see how that would necessarily help. Perhaps there is a 3x3 scramble that is just a PLL, or a 2x2 PBL. There is really no good way to draw a line here, which is why it should probably not be done. If a line is going to be drawn, I'd make it a large number of moves (14+, or something cube explorer can check quickly?), or something that guarantees the majority of the pieces will be scrambled. This leads to another line of thinking... perhaps determine how good the scramble is by how many pieces are not in their correct positions rather than how many moves it can be solved in. There are still even easily recognizable cases for this though... so perhaps ensuring the majority of pieces are out of position, combined with a large number of moves?

I had an interesting idea today about discarding scrambles. Remember the unofficial lucky criterion, where if a step is skipped <20% of the time, skipping it counts as lucky? I propose we apply the same criterion to scrambles. Basically, suppose you have a puzzle where the optimal solutions are known. Then, let N be the largest number so that fewer than 1/5 of the total number of positions of the puzzle can be solved in N or fewer moves. Any scramble with an optimal scramble of N or fewer moves should be discarded. By the way, N is 7 for the 2x2x2, 6 for the Pyraminx, 16 for the 3x3x3, and 9 twist moves for the Square-1. For the Clock it seems to me that positions with 11 or fewer moves in the scramble (out of 14) should be considered lucky. I think this is an objective criterion for when to discard a scramble or not. It also has the advantage of being more 'functional' than just discarding solves that are 1 or fewer moves away from solved, because it would also remove very easy scrambles for many of the quicker puzzles. I completely agree that single solve rankings are meaningless, but it would not be fair for someone to get an 8-turn 3x3 scramble (or whatever) and claim their unbeatable 3.xx solve. Anyone who solves an extremely easy scramble optimally will immediately get a time that they will never beat without a similarly easy scarmble.