How do you do Averages?/Average Rules

[shelley]

CCT's session averages count "completed" solves, so DNFs are not calculated into the average. This could of course lend itself to some abuse if people decide to DNF their bad solves, but it's just an at home benchmark, not an official standard of measurement or anything.

 

[cmhardw]

I've always been curious how people account for DNFs with when doing averages of 100. My turn accuracy leaves something to be desired, and it would be rare for me to string together 100 solves without having at least one jammed AUF cause me to do an incorrect turn (U coming out as R' for instance) as I put down the cube.

I personally consider one DNF in 100 to be a DNF average of 100. Perhaps I should use the standard WCA terminology and call it a mean of 100 instead. I am in favor of the mean of 100 over the trimmed average of 100. To me a mean of 100 is one of, if not the, most reliable measurements of a person's actual speed that can be done in a "reasonable" amount of time, and also in one sitting. This is why I view my record mean of 100 as a snapshot of me at my absolute best, where I define best to be an accurate representation of my true cubing skill, measured in the above defined way.

To use your example, if you DNF one or more solves going at the speed you do in a mean of 100, then that speed is not a true representation of your skill level. It is a representation of your skill when you go faster than your ability in order to try for fast 5 cube averages of 12 cube averages. However, neither of these measures your actual skill level as well as the mean of 100, in my opinion. A measure of skill is very subjective anyway, and this is the way I have always thought of my own skill level at cubing.

Chris

 

[qqwref]

To use your example, if you DNF one or more solves going at the speed you do in a mean of 100, then that speed is not a true representation of your skill level. It is a representation of your skill when you go faster than your ability in order to try for fast 5 cube averages of 12 cube averages.

I disagree. If we're talking one or a few solves out of 100, I don't think the person's going "faster than their ability" at all - speed is about how fast you can get a typical solve, not how fast you can be and still be certain to never ever mess up. Mistakes just happen, whether they're pops, huge distractions like a phone call, timer starting/stopping issues, and so on. If you do make mistakes, and almost everyone will, taking an average of 100 solves without one is actually not going to represent your real skill. Besides, how do you think going too fast is going to result in more DNFs, other than having more pops (which is a problem with the cube anyway)? It's not like people stop the timer at 10 seconds whether they've solved it or not.

 

[Dene]

To me a mean of 100 is one of, if not the, most reliable measurements of a person's actual speed that can be done in a "reasonable" amount of time, and also in one sitting. This is why I view my record mean of 100 as a snapshot of me at my absolute best, where I define best to be an accurate representation of my true cubing skill, measured in the above defined way.

You know, I hear this a lot and I have to say I disagree. I can't speak for anyone else, but I can't get through 50 most days. I would expect that I just wouldn't have the endurance to keep up 100 solves.

I think an analogy is the best way to get my view across:

When working out the body can only perform at a high level for a limited amount of time. With more reps, the body gets fatigued; muscles stop performing and mental concentration wanes. For the best results you are advised to stop before you reach this point.

Point being, when doing 100 solves, not only will your hands and arms get sore, but your concentration may not hold either. As such, I would expect weakened performance as the average gets along.

Maybe it's just because I'm unfit

 

[qqwref]

Do 1500 solves a day like Breandan and soon you will be big and strong!

 

[Robert-Y]

Maybe a stupid question but what if you just take the median if you're going to do an average of 100? Would be it better/worse than taking the mean in any way?