A “perfect” throw wouldn’t conclude anything about profile at all. You would have to introduce variations in angle like I mentioned earlier.
But no, the human throw over a relatively small sample set wouldn’t have more merit. It’s too highly variable to be meaningful at all.
That’s the point of using a machine, to eliminate the human error introduced by non-perfect throws. You’re trying to find the difference between bearings. A manual throw introduces inconsistencies in the experiment due to non-perfect throws, thus clouding the results. An experiment needs to be conducted multiple times under controlled, reproducible circumstances to validate the result. Lab testing was my career.
To clarify my tests:
I was never testing bearing shape, only perceived or claimed performance increases.
My method involved no throw or drop as these can be impacted by too may variables.
Instead, I used a rigid mount designed to hold the bearing just as a string would, and only the bearing. I then accelerated the same yo-yo to a specific RPM and released it.
This method essentially eliminates all variables except the bearing itself. It tests a bearing in real, controlled conditions on an equal playing field. Admittedly, it has a weakness in that it doesn’t account for the loss of spin due to the ‘thunk’ at the bottom of a throw… different bearings would in theory handle that impact differently… but I saw it as a fair tradeoff for the precision and control of the other variables.
To the OP’s test.
There are just -WAY- too many variables for this to be a valid data set. The throw is EVERYTHING here and it’s totally uncontrolled.
There is also a false premise -
Shaped bearings were never designed to have anything to do with spin time, only with how well a yo-yo handles strings being placed into the gap. This has nothing to do with how they spin on release, only how well they maintain their spin while tricks are being done.
There is no easy way to run this particular test in a scientific way… you would need
a controlled release
identical bearings where the ONLY variable was the shape (same company, design, tolerances, etc plus multiples of each to account for random error)
a way to introduce strings into the gap randomly (or at least in a large number of ways) and have them hold the load of the yo-yo so that you could determine how a particular shape deals with an incoming string.
Your data set would have to be pretty massive to get any kind of accurate range.
If you wanted to test your original theory of the shape impacting the spin from a throw you would only need a controlled drop (easy) and identical bearings (harder, but possible in theory)… but I’ll save you the time and say that math doesn’t support the idea.
Kyle
Variability in the throw doesn’t necessarily invalidate the data, it just increases the uncertainty of the sample mean and increases the number of trials you need to get a given level of certainty. The hypothesis test Hudakjoe used takes into account how much variance there is in the sample data, and even accounting for that, the difference in the two sample means is large enough that it’s unlikely to be explained just by the variability in the data. The experiment does give pretty good evidence that Hudakjoe’s average throw spins longer with the CT bearing than the flat bearing.
It is true that it’s hard to extrapolate that conclusion beyond the bounds of the experiment, though. The conclusions apply to Hudakjoe throwing sleepers with that yoyo and those specific bearings, but might not apply in the same way to how much spin time he gets while doing tricks, or to how much spin time another yoyoer gets, for example. It’s likely that someone with a consistently straight throw that keeps the string centered on a flat bearing will see little to no difference in this test (this is supported by the data in the test, where the two bearings have similar maximum spin times, but the flat bearing has more throws in the low end of the range due to the string sometimes contacting the wall of the yoyo). Things like variance in quality between identically branded bearings could also play a role (i.e. that specific CT bearing spins longer on Hudakjoe’s throws than that flat bearing, but a large sample of different CT and flat bearings might perform differently on average if there is variance in the quality of individual CTs or flat bearings).
Still, it’s interesting to see a school project designed around yoyos, and these kind of limitations are basically par for the course for school experiments. And it does give a nice indication of what kind of effects an inconsistent throw can have on spin times, and what kind of effect a string-centering bearing can have on that. Thanks for sharing the project, Hudakjoe.
I’m nowhere near intelligent enough to comment on the actual results. The second you got mathmatical I was completely lost.
That being said, I like the idea of people using maths and experiments to measure various aspects of yoyo performance. It would be cool to see more of it, sure seems to make for interesting discussion/debate.
Also,
I would love to see the results of this test if you still have them Kyle? 
Also, couldn’t the same rig potentially be used to measure how weight distribution affects the spin times of certain yoyos if you were to use the same bearing?
I’ve always wondered whether it would also be possible to design a rig to test “stability” of a yoyo as well. I have all these fanciful ideas and experiments in my head but I don’t have the knowledge, expertise or the equipment needed to pull any of them off. 
Valid testing or not, it was an interesting read and a good write up.
I think it would be interesting to get two identical yoyos one with a flat bearing and one with a groove of some sort, and test a bunch of throwers to see if they can tell the two apart.
I hate to be that guy but in real world when doing tricks, the player can “nudge” the yoyo back straight cancelling out the effect of tilting that eventually leads to string rubbing the inner wall killing the spin, however let’s take that aside for a moment.
The way I see it, the problem mainly lies in where you confuse chance to spin time.
Seeing the results if you ignore the low numbers, you can see the spin times are similar each other. The low numbers 30s and 40s are most likely from when the string is just incorrectly placed by chance (because it’s flat).
Maybe it’s just me, but I don’t think it’s correct to mix chance as spin time. I mean, if you have this result 10 10 10 10 0 0 10 0 10 0 seconds, that doesn’t mean it’s 6 seconds simply because it averages 6, it means 10 seconds with 60% chance, and that’s different. Your test does show that flat bearing have more chance to be “unlucky” in long sleeper with no string manipulation compared to ctx, but not necessarily prove how much or less spin time.
Not trying to discourage you of course.
You are only testing the bearing spinning in a particular yo-yo. Your test is only testing the relative spin-times of different bearings. It is unclear how the yo-yo shape will affect that.
The original question was how does bearing shape affect sleep time? The concave shape of the bearing is supposed to center the string, preventing the string from rubbing against the sides of the yo-yo. This is what we supposed to test. Dropping the yo-yo straight down will not test the specific qualities of the centering bearing that we are testing for. Only a non-perfect, human throw can do that.
Anecdotal evidence from the parts of the world where the concave bearing has not been limited by legal issues suggests that there is indeed an effect that the shape of the bearing has on yo-yo spin-time and thus performance.
Outside of the USA almost all stock-bearings are concave. Why? Because they are better; and they can.
I’ll agree to disagree.
Doesn’t make sense. If the yoyo tilts OR DOESN’T TILT (the string just really far to one side) and the string makes contact, and the player “corrects” it, it’s too late. The effect of the friction has already happened. Just because it’s not sustained owing to the player’s corrections doesn’t mean it’s not there.
But I also call a slight bit of shenanigans on this. A player is playing and going through a long and complex trick. Mid-trick there’s a bit of tilt happening… even the world’s best player isn’t correcting tilt in the middle of their fast and flowing combo. It would interrupt the pacing of the trick. They’ll wait until a good opportunity to do it.
Not to mention that correcting tilt in a reasonably fast way inherently relies on intentionally adding friction from the string to a wall. Correcting tilt doesn’t strictly require this friction (as anyone who has done the “string twisting” during sleeper contests can attest to) but to do it quickly so that you can move on, it usually does.
Yes, string nudging hurts spin time a bit; to save a lot more. On strictly long sleeper with no string manipulation vs tilt correcting (like on a trapeze), long sleeper on a trapeze is still longer because when the yoyo starts tilting, it can be corrected quickly before it goes too far, even if correcting hurts spin time a bit, this way the yoyo can sleep till it’s very slow. If you don’t correct the tilt, once the yoyo start spinning sideways, the string rubs the wall in an uncontrolled manner and it gets worse (the friction causing the yoyo to tilt even more), the yoyo dies too early.
I’m not a world’s best player but I do. It’s a lot simpler than it sounds, basically you’re simply shifting your whole tricks slightly diagonally vs the yoyo plane/direction, and it will start to tilt in certain direction. You don’t really even think about it, it’s like second nature. Like when you’re doing trapeze and you move one hand slightly forward and the other hand slightly backward, it will start tilting right? well you just swap that whole trapeze with whatever trick you’re doing and physics will do its job.
Otherwise there is no one capable of doing long horizontal combos and especially horizontal regens, because horizontal tricks require a lot of tilt correcting mid play to keep the yoyo horizontal, there is no waiting for opportunities.
Christopher Chia did eli hop while changing plane in his videos/contest routine, almost like gyro flop eli hop, if you’re interested in seeing this applied to the extreme.
I disagree
I am definitely not one of the best players out there - but for me, tilt correction and adjusting for precession is a continuous process - it’s something that happens automatically, mid-trick, right through long and complex tricks. I’ve never felt as though it was something that slowed me down or impacted on the flow of a trick - it’s just something that’s part of the art of throwing.
I have been corrected. 
Compensating for precession is something I don’t even know HOW to do unless I’m moving my body, which is a whole other story-- you’re aligning your body to the yoyo, not fixing the yoyo’s precessing.
But tilt, I’ll have to concede that it’s just my lack of skill and experience that made me imagine that it’s not something anybody would do. I still have a lot to learn!
The point about slowing down/friction still stands. It doesn’t matter if it’s for “long term gain”. That’s not the point. The point is that you’re introducing friction, period. It’s a variable that would be difficult to test for, and yet it’s one of the most significant variables in the “does shape matter?” equation… the argument is that the shape can affect how much and how often you need to correct this tilt.
Yep, quoting myself.
I was just pointing out that the shape only affects the chance of having longer or shorter sleeper, but not necessarily “how much”. A flat bearing have more chance to get a random “low spin time” occurrence than the ctx, but you can’t simply average that out and say flat bearing spins shorter.
So does shape matters on spin time?
-Does shape matters on the chance to get more spin time? yes
-Does shape matters on spin time period? there are too many variables and too little information to conclude from.
This is all interesting discussion, it helps a bit that I have some education in the subject matter.
There seems to be a lot of things that could affect the experiment. If the point of the experiment is to determine that bearing shape changes the amount of play/spin time a yoyo-er can do in one throw, then the experiment should measure several factors in order to eliminate them from the variance you are using to do your statistical test. In other words, we are gonna need a bigger model ;D
You might have to use a linear regression, or ANOVA model with several factors. Bearing shape, friction pad wear, yoyo weight, skill of the yoyo-er, as independent variables, and spin time as your dependent variable. A more advanced model can correct for variables that are going to be very hard to control.
Perhaps we are making this far too complicated. As I tried to point-out earlier, trying to control for one variable or another may risk losing the point of the exercise.
We have a large sample-size; so why not just do a simple test? Everyone who has a flat and a concave bearing - give it a try. Play your favorite, (or un-favorite, if it pleases you) yo-yo with the “stock” bearing, then try it with a different shaped bearing and see how it affects play.
A few years ago the bearing in my YYR Sleipnir went bad. I only had a flat one to replace it with. Suffice it to say that my earlier comments on this issue are rooted in that horrible experience. I have never placed a flat-bearing into a YYR since then.
It was backwards for me, I wish more people selling high quality flat bearings around here. I’m hard pressed to choose only concave/ct because there is no one selling and I had to import if I need one.
My experience is, flat bearings have less “drop” at the end of the string. Basically when the yoyo almost reaches the end of the string, it will “drop”, and you would lose control for a split second. Concaves have farther drop because it allows more strings between the response and the bearing, which amplify this problem. For casual, laid back play this doesn’t really make a difference, but when pushing it to the limit, and especially with horizontal and speed play, that split second drop gets annoying. You know that feel when a looping yoyo “stall” for a split second at the end of the string before returning? it’s pretty much like that.
Because I’m constantly moving my yoyo around layers of strings, the effect of the concave is pretty much negligible. I have been forcing myself to use concaves in the last few years because these are what comes stock with most. Only lately I tried reinstalling a flat on a yoyo that comes with concave, it reminded me how I have more control and it’s more comfortable to play.
I think spin time is only a problem when learning tricks in the sense that you have more time to think, or if you like slow-complex tricks where it matters. In the end it’s the player’s preference. For me, there is no difference in spin time, on or off trick.