Moves per second is way more important than total moves

[cubo]

TPS and pausing (...) absolutely do not improve together.

Marcel's stats prove this statement is false. Taken over the period from beginner to expert cuber, they clearly do improve together. Even if there is a stage within that overall development where one works against the other (the famous "go fast to go slow" stage), the over-arching trend is significant improvement in both.

Improving from 90s to 12s, pauses were reduced by 42.2 seconds and the time spent performing moves was reduced by 39.8 seconds. Eliminating pauses shaved more time off the solve than improving movecount and TPS combined.

Can you please tell me, showing your model, what proportion of the 39.8 second "move count + TPS improvement" you attribute to each of those factors?

It's pretty obvious TPS is mainly why we are seeing todays fast solves. The solve movecount is very similar to what it was 20 years ago, before finger-tricks and decent cubes, but the solve times are down to one third of what they were then.

Do any of the doubters about my theory want to address this observation?

 

[h2f]

Amazing! That cubo guy is still arguing about this.

And still coming up with new ways to twist the numbers to "fit" his explanation despite every single advanced cuber telling him otherwise and JanW even explaining everything perfectly in his post...

I'm not even mad. That's amazing

Agree. In my opinion cubo lost this debate many posts ago.

 

[cubo]

Anyway, the reason for the disagreement in this thread all comes down to how pauses are measured and expressed.

To my mind, solvetime = (movecount / TPS) + pauses. Where the unit of pauses is seconds. Fewer seconds paused is an improvement.

@cubo is using a model where solvetime = (movecount / TPS) * pausefactor. In his model, the 18 second solver who pauses for 9 seconds has the same pausefactor as the 100 second solver who pauses for 50 seconds and that represents no improvement.

So we can come up with plenty of examples where in the first model, and intuitively, reducing pauses is the main contributor to improvement in solve times, but when expressed as a pausefactor in the second model it isn't.

Obviously my position is that I don't agree with his model. For @cubo and anyone else who chooses to follow the pausefactor model, I can't fault his arithmetic and agree that this notional pausefactor has a smaller impact than TPS in the example he gave.

This is a nice summary of our disagreement.

I'll ask you the same question I asked JanW: could you show me, using the solvetime = (movecount / TPS) + pauses model, your breakdown of the roles of movecount and TPS in Marcel's overall time improvement?

 

[JanW]

@cubo according to your latest math the 90 second solver with 30 seconds pauses improving to 9 seconds with 3 seconds pauses does not decrease pauses at all.

 

[cubo]

@cubo according to your latest math the 90 second solver with 30 seconds pauses improving to 9 seconds with 3 seconds pauses does not decrease pauses at all.

I wouldn't say pausing hasn't decreased at all, clearly it has as a raw number, but yes I would say the pause factor is 1.33 in both cases and thus pausing contributes equally as a proportion of these total times, whereas other factors (moves and TPS, you haven't given details about the breakdown of these) contribute to the x10 overall improvement. However I would think this an unlikely scenario - it would be more likely for the speed cuber to also decrease their pause factor somewhat, as per Marcel's and Mark's actual improvement stats.

But enough about my model for a moment, since we have been over it ad nauseoum and you clearly don't agree..

Would you mind answering the question from my last post regarding your alternative explanation?

 

[JanW]

Your "pause factor" is a completely meaningless stat. A 5 second pause is a 5 second pause. It takes just as long no matter how fast your solve is. Eliminating that pause saves as much time regardless of how fast your solve is.

Would you mind answering the question from my last post regarding your alternative explanation?

What about it? For the last layer it's 50/50 as was established a long time ago. Calculating TPS for the rest of the solve is virtually impossible.

 

[mark49152]

I'll ask you the same question I asked JanW: could you show me, using the solvetime = (movecount / TPS) + pauses model, your breakdown of the roles of movecount and TPS in Marcel's overall time improvement?

No, partly because I don't have time today, and partly because I don't think comparing those two solves is meaningful (sorry @JanW). One is so slow as to make it impossible to discern between pauses and low TPS, and the other is a 12 second outlier from an 18 second solver so is unlikely to tell you anything more than what want exceptionally well on that one stellar solve (I would bet that 0.7 seconds total pauses is not the norm, sorry Marcel ). Instead I offered you my own improvement data, based on averages and splits where I had enough information available to make sensible estimates about the three metrics.

I hope you have found the inputs on this thread enlightening, but I suspect that nothing I or anyone else can say here today will deflect you from your own certainty in your own conclusion. My suggestion would be that you go practise for a few months and come back when your times have halved, and then maybe you'll have some different insights, and be better able to understand what we're saying, having experienced it for yourself.

 

[cubo]

Your "pause factor" is a completely meaningless stat. A 5 second pause is a 5 second pause.

Actually I've shown it's a mathematically consistent way to represent the time effect of that factor on the solve, and furthermore, can be used with the other factors to compare their relative contribution to time improvement.

I must say I'm surprised you think pausing can be magically reduced without any effect on the other elements. Your model assumes that a 100 sec, 100 move, 50% pausing beginner could, if they simply focused on look ahead and reducing pausing, become a 50 sec, 100 move, 0% pausing cuber. This goes against the evidence of both Marcel and Mark's stats, which is that all 3 factors improve together, at varying rates.

Lastly, "a 5 second pause is a 5 second pause" is as meaningless as stats get. X = X? No kidding.

[Edit to add final thoughts ...]

3rd time lucky? Does anyone care to explain either of these without admitting TPS is the crucial factor?

It's pretty obvious TPS is mainly why we are seeing todays fast solves. The solve movecount is very similar to what it was 20 years ago, before finger-tricks and decent cubes, but the solve times are down to one third of what they were then.

I think this is an interesting video that helps to prove the point that a fast TPS can be more effective than move count. As he manages to get sub-20 using the beginner method.

Actually don't bother because ...

I hope you have found the inputs on this thread enlightening, but I suspect that nothing I or anyone else can say here today will deflect you from your own certainty in your own conclusion. My suggestion would be that you go practise for a few months and come back when your times have halved, and then maybe you'll have some different insights, and be better able to understand what we're saying, having experienced it for yourself.

Thanks @mark49152, I'll happily take your advice. It has been enlightening, even if I've been almost entirely swimming against the current on this one.

Same goes to everyone else (@JanW, @AlphaSheep, @shadowslice e, @Escher, and all the others) ... thanks, and bye for now.

 

[tseitsei]

This is getting ridiculous already, but I'll try one last time.

The reason I won't calculate my "breakdowns for MarcelP's solves is that in the longer solve it is quite impossible to determine his TPS during F2L since he literally pauses after every single move. However even during the last layer he improves as much (or more) by decreasing pauses than by increasing TPS and reducing movecount. And everybody (should obviously) understand that during the F2L is even much more important than during last layer. Last layer is the portion of the CFOP solve where you just spam algs with maximum TPS as opposed to F2L that you often solve intuitively and with lots and lots of lookahead to reduce pauses.

Actually I've shown it's a mathematically consistent way to represent the time effect of that factor on the solve, and furthermore, can be used with the other factors to compare their relative contribution to time improvement.

I can show many things that are mathematically consistent but don't really make any sense in the reaö world.

I must say I'm surprised you think pausing can be magically reduced without any effect on the other elements. Your model assumes that a 100 sec, 100 move, 50% pausing beginner could, if they simply focused on look ahead and reducing pausing, become a 50 sec, 100 move, 0% pausing cuber. This goes against the evidence of both Marcel and Mark's stats, which is that all 3 factors improve together, at varying rates.

Of course pauses aren't magically reduced just as movecount doesn't magically drop or TPS doesn't magically increase. However it makes much more sense to use an actual and absolute number for pausing rather than some arbitrary pausing factor. Since 5s pause in a 100s solve is just as long (5 seconds) as it is in a 20s solve (still 5 seconds).

Lastly, "a 5 second pause is a 5 second pause" is as meaningless as stats get. X = X? No kidding.

See the last phrase of the above chapter. Athough I have to say that you really are the master of meaningless stats based on what I have seen in this topic.


Anyway this will proably be my last post in this thread since you have obviously decided how things are and this can't be changed by anything.

P.S. oh almost forgot this:

It's pretty obvious TPS is mainly why we are seeing todays fast solves. The solve movecount is very similar to what it was 20 years ago, before finger-tricks and decent cubes, but the solve times are down to one third of what they were then.

Do any of the doubters about my theory want to address this observation?

How many people do you think actually practised speedsolving seriously 20 years ago?
That number has gone up quite a lot so that is the main reason solve times have gone down.
With more people solving you get more talented individuals and also good tutorials so that it is easier to get faster.

Yes I agree that improved hardware has played an important part here too. But the most important thing is that the whole cubing scene has grown pretty much exponentially.

Also as I have said before. Once you get ~sub-15 TPS starts to be more important since you have already reduced a lot of pausing so increases in TPS are needed to get faster. But going from 100s solver to 10s solver by far the most time will be saved by eliminating those pauses. Going from 15s solver to a 10s solver TPS might be important (even though I still think it's lookahead and no pauses but at least TPS has a bigger role than when going from 100s --> 10s).

 

[sotolf2]

Yes I agree that improved hardware has played an important part here too. But the most important thing is that the whole cubing scene has grown pretty much exponentially.

A proof of this is that top level solvers can get sub 10 times even on old hardware

Also to compare most of my solves (Averageing around 50) and slow solves of @PenguinsDontFly (sub 15s slow solves) My TPS is higher than his, but my pausing is crazy, since I still suck at lookahead. The pauses are what's keeping me down, not my TPS

 

[shadowslice e]

If that's what you're doing, then I think a better title for this thread would be: "Tps improves the most in transition from beginner to world standard" or similar. I'm sure you could think of a better title in the same vein.

Just going to reiterate this because I feel like if you just changed the thread title to this or similar it would absolve you of most of your detractors.

 

[mark49152]

It's pretty obvious TPS is mainly why we are seeing todays fast solves. The solve movecount is very similar to what it was 20 years ago, before finger-tricks and decent cubes, but the solve times are down to one third of what they were then.

Do any of the doubters about my theory want to address this observation?

I'm pretty much done with this thread, but I can't resist the temptation to suggest that 20 years ago they had the same TPS but paused for 2/3rds of the solve . That would explain this observation indeed. Obviously that's not true, but @SenorJuan's observation clearly is only considering overall average TPS for the solve without accounting for max TPS and pauses.

Anyway, @cubo, this is a friendly place and you're always welcome here so don't be disheartened by the disagreement. Just remember that some people have been doing this for many years, and if you come in expounding theories that go against the consensus built from years of experience it will of course feel like swimming against the current. One of the things I like about this community is the willingness to share advice and help each other and of course all the advice offered here is in positive spirit.

One last thing. Since you're interested in analysing performance and modelling improvement, this thread might appeal to you: https://www.speedsolving.com/forum/...aining-that-yields-systematic-progress.39406/

 

[cubo]

Thanks @mark49152 I agree this is a great community with a great vibe and positive spirit. I love to challenge the status quo but I don't ever want to be disrespectful to the amazing cubers who hang out here. I'll definitely check out that other thread, thanks again!

 

[AlphaSheep]

It's been a fun thread and I've enjoyed reading the replies. Just one last point I need to make.

Actually I've shown it's a mathematically consistent way to represent the time effect of that factor on the solve

The thing I love most about maths is how its always consistent, and you can derive almost anything from it. The thing that's most frustrating about maths is it provides absolutely no way of telling if your conclusions are physically meaningful or not.

I remember practically all of my lecturers at university constantly repeating to us to always step back and check whether or not our answers actually make sense in the real world.

Yes, you can divide the pauses by the total solve time. Yes, you get a factor that can be compared to other factors. And yes, you can say that one factor is greater than another. But thats as far as mathematics can take you. The conclusions you then draw from those comparisons are not mathematics. They are based on interpretation, and that interpretation needs to be grounded in experience gained in the real world.

 

[the_chad]

Ruihang Xu took this thread too seriously

64/3.77=16.98TPS