Marble size sorting machine

[@thomaskaldahl196]

its really satisfying how the marbles appear to slow down before they drop, as if each marble has a final destination in mind and slows down in anticipation for its arrival in its new home

[@bigabong7119]

Thomas Kaldahl dude. This thing needs more likes.

[@caseykoons]

I like it when you get mathematical. Thanks for your hard work.

[@pocket83]

Thanks for the appreciation!

[@RedScaledKnight1]

it's great to watch. he's also got great mechanical advice ;)

[@Jer_Schmidt]

The way they brake smoothly to a (near) stop before dropping is so satisfying :) Really cool device!

[@deeredmond8285]

Being creative and logical is a curse, but watching you use both sides of your brain at the same time... WONDERFUL (I don't feel so alone now) I bet you drive other people crazy too, so much more to say but jumbles up... Thanks for the inspiration and a fresh breath of air

[@earthbjornnahkaimurrao9542]

I think it is neat that the marbles automatically slow down right before falling through.

[@pocket83]

No they don't! They're just following a curved path. Along and then down, which is longer. They are actually speeding up- watch the spin. Mindblow.

[@KainYusanagi]

They still slow down with regards to horizontal movement, though, though they do not slow down entirely. ;P

[@pocket83]

I guess you're right; they slow down rapidly with respect to that. In fact, they stop making horizontal progress entirely once they hit the bin.

[@KainYusanagi]

Haha. I was referring to the frame of reference of them on the sorting slide, as Earthbjorn stated, but that's true as well.

[@Keldor314]

To see what's happening here, think of the part of the marble in contact with the tracks. It will roll on circles near the end of each side of the marble, right over the point of contact. As the bars move further apart, the circles shrink, since they get closer and closer to the ends of the marble, until finally they shrink away to nothing and the marble falls through. I diagram would be useful here, but I don't have one :-P Anyway, the smaller the circles of rolling contact get, the faster the marble must spin to move at the same speed, just like a small wheel has to spin more than a big wheel to cover the same ground. Since friction is keeping the marble from slipping as it goes down the track, it has to spin faster and faster. At the end, right as the circle shrinks to nothing, the marble would have to spin infinitely fast, but of course it starts slipping eventually. Still, this is why the marble slows down before it falls - more and more of the momentum from rolling downhill is pulled in to angular momentum to make the marble spin faster and faster as the circle of contact shrinks away, until the marble very nearly stops right before it falls through. The higher the friction between the marble and the track, the closer it will come to stopping. To demonstrate this, it would be cool to run some rubber balls down the track ;-)

[@lank_asif]

pocket83 When you adjust the bars to further sort the selected "average sized marbles" you should take one of the smallest outlying marbles and set the top of the sorting device to the width of that marble. Now you'll be sorting the marbles more finely over the entire length of the device and should more closely recreate the theoretical curve. As it stands, the sorting only begins at about a quarter of the way down and so your results will be less accurate. This is just my suggestion. as I love the concept and your content is fantastic.

[@UPWNU]

Awesome approach to something that was nothing but laborious to learn in my AP Statistics class in high school. It's always a great pleasure when you do maths in your videos. :)

[@AndreaArzensek]

I could watch that marble sorting all day long, simple - yet fascinating!

[@pocket83]

Ha! I know what you mean. I found myself staring at it for a long time. That's why I left the opening footage so long.

[@mattjohnston2]

Hey, it's good to see a vid! It's funny...I have absolutely no intention of making a marble sorter. No need, no desire. But I thoroughly enjoyed the video! It's almost like sitting down with a friend and having a conversation about one of HIS interests. I don't have to have any "take away" from a conversation like that to enjoy it. So...thanks! I also appreciate the metric conversion ;)

[@AndreaArzensek]

I just watched How it's made on Tinned Pineapple and guess which method they use to sort the pineapples to sizes :) I instantly remembered your video! Cheers!

[@steveschierholz5272]

You did a good job. The way you sorted the marbles by color was amazing. So they got sorted twice.

[@jordanengdahl4418]

Perhaps it is beyond the scope of your expertise, but a video on carburetors would be very interesting. Such a delightful little mechanism in a small, inexpensive and mechanical package.

[@Makebuildmodify]

You must have presorted the marbles in the beginning of the video. Right? It really had me thinking. There are other possibilities that would cause the same separation to occur, but I would assume them to be more unlikely.

[@pocket83]

Yep.

[@zachell1991]

I was wondering the same thing.

[@lance00000000000000]

zachell1991 same

[@TimothyHall13]

I really like this. We human beings do have a hard time adjusting to statistical concepts and this does a great job of illustrating distributions. I imagine if that you would have set the rails by choosing the largest of the large group and the smallest of the small group you might have ended up with an even more normal looking distribution? Thanks Pocket!

[@dmoore3722]

Your videos are so inspirational and thought-provoking, thanks for sharing!

[@pocket83]

Thanks for letting me know that you like it!

[@danno1111]

Could very well be that the distribution isn't normal, depends entirely on the method that the marbles are made. If they rolled the molten glass between two steel rollers to form the sphere, there's a set upper boundary to the size, and the distribution would fall off from there in the smaller sizes, depending on the temperature of the glass or something else. Not to say that that's how they do it or that there isn't a normal distribution, just that that's a possibility.

[@pocket83]

Good point. Similarly, when I adjusted the lower part of the device (and not the top), I only extended the variance of the new set in one direction, larger; there was, in this case, also a boundary... wait... technically, that's not right- since the parallel bars had a new, completely different slope after the adjustment. Who knows? Though provoking stuff, isn't it?

[@MrAsylumEscapee]

Think you are correct that you only extend the variance in one direction: by adjusting the lower part of the device only, you apply a linear transformation to the set, with width at the very bottom changing by X, and width at the very top not changing at all. This would skew the entire distribution to the left, since all marbles will fall off later. The variance will increase, but strictly in one direction. (Since marbles cannon physically fall off sooner).

[@MacJohnsonFarm]

I came here looking for inspiration for a DIY machine to sort nuts by size. Fascinating watch. Thank you.

[@pocket83]

Thanks. That's a much harder problem. The first idea to spring to mind is a rotating, vibrating drum that has thickness-based slots which would allow smaller nuts to escape. Might need some oil. Same theme as this video: over time, math gets closer and closer; eventually, all the small nuts _will_ escape! I'd be interested to hear what you come up with.

[@randyt]

Another great video! It would be interesting to see this incorporated into a marble machine as a way to direct marbles onto different paths.

[@pocket83]

That would be spectacular. Imagine a double helix using two different colors. It would be mesmerizing to watch it self-sort and then spiral down around itself.

[@ComplexVariables]

What a wonderful demonstration for a statistical reasoning class!! Will keep you updated if I can find a makerspace student to make one and then try it in class.

[@pocket83]

Great. I'm also glad that you 'get' the big picture that it's a useful metaphor; I've received some criticism from statistics nerds over it not producing a _true_ Gaussian distribution. As if I could be expected to use an infinite number of compartments. I think this would be great for discussion in a class.

[@ComplexVariables]

What in RL is actually a normal distribution?? I totally got that you were just pointing out a very typical distributional shape. I suspect marble makers have an idea of a "good" theoretical distribution model considering the manufacturing process (some kind of Gamma distribution perhaps ) Anyway, I am amazed about the sensitivity the results are to the deflection the marbles placed on the steel rods, a nice example of noise reduction.

[@PAFrogBoy]

The first year engineering class at my university assigns a semester-long group project, where we have to design a marble sorter and barcode reader out of a Lego Mindstorms kit, which has some motors, and various types of sensors such as touch or light/color sensors, and program it with LabVIEW. The finished product should be able to read a barcode containing a prescription of certain size/color marbles to collect, and fill out the dosage from a pool of about 75 marbles. It was a very challenging project and this video made me think back to it and how crappy our design was

[@Nevir202]

I think a sorting system like this would be a really cool feature on a very large marble machine. You could have several separate tracks which each use a single color of marble, then they all merge to go through the last section together, and then as they are lifted back to the beginning of the course, have an array of these sorting them back out so that they continue to go down the course with the same color marbles again and again.

[@KainYusanagi]

I know this may sound weird, but... could you do a video of you just sorting all your marbles using this machine? The tink of the marble on metal, the slight sound as it rolls, the tap of it hitting the wood then another marble, it's all just so soothing.

[@pocket83]

It's not that weird! There's lots of videos like that out there.

[@KainYusanagi]

I meant weird in the context of asking you specifically to do it, but yeah, that's actually exactly why I asked.

[@Checkedbox]

You have the best sounds, the audio is so clear and crisp

[@pocket83]

Thanks for saying so. I often wonder about that. Usually, if they aren't complaining, I don't worry about it.

[@RobertMilesAI]

How much variance is there in the process? If you put the same marble through 100 times, what's the spread on where they fall through?

[@pocket83]

Marbles each have their own personality. The more egg-shaped ones will give a wider spread. In general, consumer marbles with a reasonably high tolerance will fall into the same cell 7 or 8 times out of ten.* This is a great question, because in projects where precision marbles are needed (yes, that's a thing for me), variance is a key consideration in the marbles that I select. My usual method for establishing a tolerance goes something like, _this set must land in cell x at least five out of six times._ Variance has everything to do with roundness. *Note that this is largely dependent on the distributor. In my experience, craft stores have the least consistent quality marbles, and, perhaps ironically, dollar stores usually have flawless gems mixed into their $1 per 100 ct 'variety' bags. Seriously--I've found like nearly optical quality marbles there in at least one case.

[@RobertMilesAI]

Ah it makes a lot of sense that it's about roundness. I guess it might end up depending a lot on the orientation it's placed down in. Like, it might end up rolling around its major axis and seem bigger, or around its minor axis and seem smaller, or somewhere in between. But there's no easy way I can think of to control for that

[@BrebtaGamesSK]

You should put a funel on the start where you can just put many marbles at once

[@GretgorPooper]

There's something so satisfying about the sound of marbles falling on wood.

[@billyvillacis9975]

This will prove invaluable in my upcoming marble line-breeding experiment!

[@earthbjornnahkaimurrao9542]

I'd like to see the same thing but with smaller bins to get finer resolution on sorting and make a better distribution "graph".

[@Dapstart]

This is the quality content I love seeing from you. Brilliant ideas as always :D

[@MrBlahChannel]

I'm glad you're back. Comfy video too

[@pocket83]

Thanks ;)

[@misoman]

I could watch those marbles all day. I hope we're getting a pocket squared video on this later, wink wink

[@genghiskhan6688]

It is so satisfying to watch those marbles rolling ...plz post a video of just that

[@Newmachinist]

Thanks for the great videos - really interesting and educational. Thanks Rod

[@ian-duh]

I'm just more impressed that it's the same colors in the same categories at first.

[@Mike_Hoffmann]

Damn! Nice unit. Now I need to find my marbles.

[@PinBox3000]

The reason the marbles slow down before they drop is very easy to explain: Magic.

[@custersword7746]

Very cool machine for sorting and statistics.

[@MrUltramove]

Semi automatic loading mechanism would be great for this, so you can let each marble sort without bumping into each other but increase sorting speed for big sets of marbles. Just some loading chamber, channel to line marbles and pushing leaver with rubber band to push it back to lock position. Hope you dont mind this input, love your idea, your loyal fan.

[@pocket83]

That would be cool. It would also be fun. I wanted to make it as simple as possible though.

[@alfredneumann4692]

In the Gauss-Distribution modell all marbles have the same diameter and go all through the same opening. Then they fall down through a bunch of nails (or similar). I have build this years ago for a demonstration in an institut in a University in Germany. The plan was in a mathematicsbook. Not really, but i used the schematic and build a 1m (about 3 feet?) model. Was a nice project with many try and error. Cheerio from Germany

[@pocket83]

That sounds like the original Pachinko machines. It can be used for gambling. In that case, you can have jackpots on the tail-ends of the curve, and the 'house' still wins. Note: I don't gamble. Ever.

[@nickm5911]

Another amazing video! Awesome job!

[@jin3784]

In an earlier video, you stated you wouldn't be a good teacher in a classroom because you get to edit out your mistakes, but this video demonstrates that you could be a very helpful *online* teacher! Y'know, maybe something along the lines of Khan Academy, or a Pearson example video.

[@feygilo8601]

So glad your back!

[@HKKyoya]

Watching the marbles get sorted is eye candy. Now what if you turned it into a display? Where the marbles get automatically re-fed into the top. I think I'd never get any work done...

[@saartal4524]

Nice design. Another option (crude) is using sifts of various sizes - which I have made Thanks for sharing

[@azimovwatts6425]

for some reason this reminds me of the "wave resonance" produced in the single particle double slit experiments.

[@markfoxwell79]

Very interesting. Have you seen how ping pong balls are spherically graded? They are rolled down a wide straight ramp towards 5 collection areas - the balls that go through the centre collection area (least deviance from a straight line of release) are graded 3*, the 2 areas either side of that are 1* and the rest are ungraded balls. I think this could be easily replicated for any projects where roundness in key.

[@pocket83]

Wow, that is really clever! But what if a malformed blob came out? It would just slide down the center of the ramp ;( Note: grade -1* are probably used specially for carnival games!

[@Lemongrasspicker]

That is clever, nice work sir!

[@Wordsnwood]

The opening 45 seconds was OH SO SATISFYING to an OCD'ish personality. :-)

[@darfjono]

i saw some marbles push other marbles further down, is that a major problem?

[@pocket83]

Not often. The weird ones do jam the device sometimes. But you have to expect some anomalies, deformities, and unlikely events. That's the nature of it! You can always repeat the run to get better results.

[@nicholaslau3194]

Do one marble at a time. Only put the next one in once you see the previous one has fallen

[@pocket83]

Agreed. But patience is easier said than done!

[@KainYusanagi]

You could add on a funnel system with a rotary paddle release mechanism at a certain tension so it only releases marbles as a steady drop rate, so no two marbles will interact on the guide, no?

[@White-ul3dl]

pocket83 why seperate by size when they are all sizes are different colours?

[@Gunbudder]

you are graphing the random errors introduced during manufacturing and it is pretty cool. theoretically, if there was no random error, they would all be in the same cell. also, if the difference in diameter was deliberate, the shape of the curve might be different. very interesting. i don't like your pile of sand analogy though because the shape of the sand has to do with the particle size of the grains, and gravity (and not randomness).

[@pocket83]

A pile of sand _is_ a distribution. It has both a center of mass and a geometric center, which are types of averages. The forces that provide the variables for its measurable characteristics are incidental. I stand by the analogy ;) And there is no such thing as random!

[@Gunbudder]

pocket83 ok, fair enough. It's a good analogy :). Thanks for the video btw

[@jwrm22]

Sorting marbles can be a start of a analog computer. Each marble representing an instruction. Ran trough the 'computer' one at the time.

[@hiimapop7755]

Ayy Its been a while! Good to see you upload a new video again!

[@pocket83]

Thanks! I'm also working on a build video, so expect some more soon ;)

[@trentw26]

This is awesome! What possible reason could someone have for disliking the video.

[@pocket83]

I think it's more of a case of them disliking _me._ I mean, no real human being dislikes watching pretty rolling marbles organize themselves, unless they've gone completely deranged.

[@trentw26]

Well as far as them being deranged, I concur. One might say, they have lost their marbles.

[@b5a5m5]

I wonder what sort of tolerances can be achieved with this type of sorter. I've been playing with the idea of using this to sort airsoft bbs for use in 3d printed bearings. I imagine the precision on the initial rods is what would be the main factor in the machines sorting capability coupled with the rigidity of those rods. Maybe some long precision ground shoulder bolts if there are any long enough... Oh! Perhaps linear bearing rods would work well!

[@pocket83]

Look for an old printer. The rods inside are flawlessly smooth and precise. With respect to tolerance, mine uses inexpensive zinc-plated rod, and it goes way beyond my visual acuity. If you really _need_ to know a numerical value, get back to me and I'll take some measurements. Good luck.

[@kosdan]

What surprises me is that the rods are straight enough that it doesn't affect the device. I wouldn't think that they would come so straight. If any rods was just the slightest bit bent or warped or changed thickness you would see the distribution be incorrectly skewed.

[@pocket83]

Good point. The zinc coated rods did come from the store unusually straight. However, even if they weren't, the device would still work so long as no adjustments were made between distributions. A distribution would still be in there, even if it were skewed a bit in places, so you can still find a subset of marbles that share a particular size. You just conjured the picture of me looking at a Bell curve through a glass-block window; the curve is continuous, but it would have local maxes and minimums. I want to add that a rod can also be bend-straightened to a pretty high tolerance just by sighting down it and using your eye, especially in the case of thick rods like the 5/15" ones used here. I also want to add that the device here was for sorting marbles to a reasonable standard: so while it's fun to question, speculate, and critique, it's also somewhat silly to think too far into such a novelty. lol. I struggle with that sometimes. But really, the thing does have a certain (unavoidable) amount of error. Supreme precision and accuracy would likely involve fancy optics, lasers, and other tech that is squarely out of reach for a troglodytic carpenter such as myself.

[@kosdan]

Thanks for taking the time to reply and share your insights. I liked your glass-block window analogy.

[@LINGLING-ep8ev]

shouldn't be too hard having a funnel above the machine allowing you to dump a bunch of marbles and let it do its thing? your genius enough!

[@SethPortland]

Thanks now I can tell my senile grandma how to get her marbles sorted.

[@MrMagicBlox]

What if you made it so that the bars were movable (within a guide of course) and you set values to the compartments on the tray. Then give players marbles and have them try to get the highest score within 3 moves. I imagine that the coordination required to move the ramp to keep the marble on the "ramp" would be challenging. The higher scores/values would be closer to the top. Play would start the marble at the bottom and move the rails to get the marble closer to them. The physics of this would make an interesting video in itself.

[@bubblegumcat6822]

I apologise for this comment being irrelevant to this specific video, but I remember a video you made a while ago with interesting imagery and nice cinematography, and I looked though both of your channels without a sign of it. It's a little disappointing, since I wished to see it again but I respect your choice to delete it. Your videos are great, and I'm glad that I came across you when I did.

[@pocket83]

Thanks. What was the premise of the video?

[@bubblegumcat6822]

pocket83 I don't remember things well, but I know it had multiple shots of the sky (I believe from a swing), a grasshopper/cricket in a jar and other similar things. It sounds silly, I know, but I found it very interesting and strangely soothing.

[@pocket83]

That was last year's wildly unpopular Halloween video. It's now unlisted. Here's the link: kzbin.info/www/bejne/Z5mkn4Rtm9iGmrM

[@Handsy_McGee]

Just saw an ad for "Tumblestix". You called it in that kururin video, man.

[@mversantvoort]

Simple yet very cool!

[@charliecollard6433]

Neat how they slow down before dropping, ....!

[@Keldor314]

One thing about marbles is that they're generally not very perfect spheres. It would be interesting to take all the marbles that fall into one bin and run them through the sorter, unaltered, again, to see if they still land in the same bin. Incidentally, when they sell ball bearings, they have grades according to how perfectly spherical they are. Found this out when I needed to get 3/8th inch balls for a marble machine.

[@pocket83]

Repeated trials *will* increase precision. There are certainly defects, but manufacturers are aiming at spherical in general, so they often are quite close. Still, there's a reasonable degree of quality that you will try to achieve with this gizmo, that is, a spot that rests within a certain variance. Once that sweet spot is found, the marbles that hit the same bin a few time in a row are extremely spherical; variations and anomalies will readily register as a different size occasionally. All things considered, I'm more impressed by the sphericity that marble makers have achieved than their ability to sort within a tight size tolerance- not that most kids care what size they are while they are rolling the things around on the porch.

[@homer2336]

Wish I could use this concept to sort laundry.

[@jkell42]

As I am watching, I wonder why not make a hopper?

[@pocket83]

Simplicity.

[@Jaynautic]

This could probably replace manual sorting labor with using electricity (if the system is refined).

[@toushevtsurpe8568]

it's simple BUT AWESOME!

[@Alphabetac]

"okay okay i dont want to bore you..." I literally came here to listen to the sound of the marbles being sorted

[@spokehedz]

I think (this is all theoretical) if you wanted to get your marbles into sizes that you actually want, and not have to adjust the rig many many times, you could use a known size to set the bars apart. Or, if you were moving from something with very precise tolerances (Ball bearings) to lower tolerance (Ebay/dollarstore marbles) you could set the bars for that known, working size. So if you were designing something around, say, pachinko balls... But they are more expensive, and heavy since they are metal. So whatever the reason, you need to switch from them to marbles. Since the metal ball bearings are mass produced and require tight tolerances to work in the game, but marbles are not, you would know what size you want (11mm). So, by putting the top of the bars 10mm apart, and the bottom 12mm apart you could get your initial and highly refined sort very quickly. You would probably have to modify the design to kick out the super large marbles on the end, as you would probably have lots of them because of the initial sort being very precise. Maybe I'm over-thinking it...

[@nbayoungboyrichtvgaming676]

Wow nice marbels

[@lance00000000000000]

So why dont marble making companies have this sorter? I guess it's because its not super important for most people who buy marbles but idk. Also the first run you can not find much of a slope but if you re run each section of the first run in order and add them in it would be more accurate and another better way would be to lower the decline on both the rails lower and collecting system plus lower the height of the drop then instead of a few dividers you could have alot of ridges to hold them in place with an average marbles distance between them so they are all in order basically from smallest to largest. The lessened height would just make it so the marbles wont bounce over the ridges or roll too far down the rails. Anyway very interesting video, thanks for sharing this with everyone! Edit: the second option would give a much better visualization of the average size but is passed the whole point of the size sorter as it would not be as simple to group them together because you would end up with many groups, which i guess would only be better for extremely precise projects

[@zaxonov]

To avoid the bending problem while adding a lot of precision of the sorting process, I imagine a multiple pair of bars builded like the original Donkey Kong game :D However, fine tuning this sorter could be tricky I think...

[@초초-l5g]

Can you make a funnel that lets marbles pass through with a window gap of 1 second?

[@hovissimo]

I didn't see you address repeatability. I imagine most marbles aren't all that spherical, and so the same marble might fall into different bins on consecutive runs when your bars start approaching parallel. This might actually be a good technique to measure sphericity in your marbles too! (A more consistent binning at very tight tolerances suggests a more spherical marble)

[@pocket83]

Yes! Exactly.

[@Cumbernauld]

Upload more your vids are awesome

[@pocket83]

I'll try to have one next week. This stuff isn't easy to make, so bear with me!

[@randallmcclain6368]

Outstanding!!

[@homunkoloss6782]

This reminds me of the Marble Machine (x) that wintergatan is building right now

[@milkproductions8127]

Have you tried a hopper mechanism so you won’t have to do it one by one?

[@zacharynovkov-bloom4933]

This is really cool, love your content! Could you try making some sort of web shooter using the two liquids 1 6 hexanediamine and sebacoyl chloride?

[@pocket83]

No. It smells too bad. And it burns.

[@Dman0001]

Pocket should be a math teacher. He's also head and shoulders above bill Nye as a scientist

[@KagemandenDK]

Can you make this much bigger to sort cans and tennisballs?

[@parththummar9110]

Hello @pocket83 Very nice idea. Does this machine work for sorting of balls of size 10.1 mm and 10.2 mm?

[@pocket83squared]

Of course. You may need to modify some dimensions, but yes, the basic premise will work. My machine can easily detect size discrepancies of 0.5 mm (on average). You can increase the machine's sensitivity (precision) in two ways: 1) by moving its bars closer to parallel, and/or 2) by reducing the distance between the collection bins. Also note that the effectiveness (accuracy) of the machine can be increased by running multiple trials. The more perfectly spherical the balls are that you're sorting, the fewer trials you will need in order to accurately sort them. Marbles are really imperfect, so they take a few trials.

[@parththummar9110]

@@pocket83squared Okay, Will do that and try. Thanks a lot for reply.

[@Majoofi]

Are different color marbles ways of coding the sizes? I was surprised to see that in the first batch it sorted colors exactly.

[@pocket83]

Not exactly. Different styles usually have different averages, since they have a different history. But those averages overlap, since most marbles are around 9/16". The opening shot was made by using pre-sorted sets.

[@AlexJohnson-un5mb]

could you build some sort of hopper to feed the marbles in? to save you having to place them on one by one.

[@imp3r1alx]

wow.. i like how they kinda slow down for a bit before falling down.. as if they know it's their stop =D

[@jeffadams9736]

Is it an optical illusion, or do the marbles reverse rotation as they drop between the rails? I cannot think of a reason they would, but they appear to.

[@pocket83]

They do not. Ceiling fans and car wheels often do something similar when video recorded. It has to do with a pattern that emerges from capturing a specific number of frames per second. Your brain puts what it sees together wrong! A similar cool phenomenon is the Moire effect. Screen doors do it. Here- I made a video a few years ago: kzbin.info/www/bejne/h4u8oXiDpZ1qjtE

[@Sludgepump]

Fun to watch the rpms go up as the bearing surfaces dia decreases.

[@JeremyCook]

Would be such a good teaching tool for statistics. Now, how about a device that illustrated Pareto's Principle? Only like 25% joking, I think that would be awesome, though have no idea how it could be done.

[@iCONAN1]

That’s neat! 😃

[@CthulhusDream]

It's all well and good as long as you don't lose your marbles.

[@pocket83]

Ohmygoodness, I've learned where that expression comes from. This machine dropped another marble under my planer every three minutes. I'm working on the prototype (that this relates to) on the carpet.

[@andrewmalaty8]

You should make a hopper or funnel so you can put all the marbles in at once.

[@ihavecancerandaids2737]

This man needs to become a teacher

[@honormarie13]

Where did you buy your marbles at ? Could you add a link?

[@YostPeter]

Can you build a machine that helps me find my marbles? Thanks.

[@ChrisAurora]

I have recently been thinking about this channel whenever @wintergatan has uploaded more videos from the building of Marble Machine X - and even more so now!

[@stanleydenning]

I see possibilities... A multi run marble machine. Each run for a slightly different size marble. All starting on a single run and than being redistributed by size, than combined back together again just before the lift.

[@RCWorks]

Now you should make a auto-feeder so you don't have to load by hand.

[@justwondering5540]

OMG where are we going with this ?I'll have to cut more paint cans open to get marbles lol

[@rosseatssnakes]

Why have you not been working with Wintergaten!?

[@NathanielMitchellnm]

... And what is your production budget? Kind of hard to consider if we don't know what it is. ;)

[@pocket83]

Not much.

[@Keldor314]

Let's see, we already have $6 for the bars (more if you count the ones that didn't quite work), then throw in some nails and a piece of plywood or two to make things out of and... Probably less than $20, less if you already have some wood laying around.