OOAK Table Tennis Forum

Hi all,

I'm currently in the market for a robot and to be honest, I'm waiting for the new Amicus to be made available here in the UK before I buy anything. However a friend of mine kindly leant me his robot so that I could test it out on my table at home. This was mainly to see if I had enough space to play in my garage with a robot and also to see if I enjoyed using one as a practice tool.

My friend’s robot is not a make that I know. He said it was quite cheap (circa £300) and is a Chinese brand (A Super Master robot - photo below). For clarity the robot has two wheels (top and bottom) to generate the topspin and backspin. Each wheel has speed settings from 0-9 (0 means the wheel is turned off and 9 is it’s running at its fastest).

Now the crux of my problem is that I think it might be throwing my game off. After setting it all up, I've found it difficult to get it to produce a ball that bounces naturally. In other words, the bounce feels unrealistic.

To give you an example, if I want a top-spin ball, I set the speed of the top wheel to say 3 and bottom wheel to 0, the ball just fires into the net. In order to get the ball to go over the net, I have to have the bottom wheel moving as well (setting it to a minimum speed of 1). This means that the ball produced has less top-spin on it as the bottom wheel is counter-acting the speed of the top wheel and thus, it feels as if the robot is producing more of a float ball. To get around this, I can of course turn up the speed of the top wheel to produce more top-spin, but this just makes the balls fire faster at me, rather than producing a slow top-spin shot.

After practicing with this robot for a weekend, I played a league match the following week and I really struggled to get the ball on the table as my timing was completely out. I can only put this down to the fact that the robot has been giving me float (or unrealistic) balls and when I warmed up against a real player, I was over-hitting the ball and my timing was completely out.

So really my question is, do all robots give this unnatural feel with the bounce of the ball or is this a symptom of using a cheaper robot? In other words, does a more expensive (better made) robot fire out balls that bounce as if they have been hit by a real person, or is this artificial feeling just part of using a robot?

The bottom line is that I don't want to invest a lot of money in something like the Amicus if all it's going to do is throw out my timing and ruin my game. The whole point of getting a robot was to practice when I didn't have a partner to play against, but it seems pointless if it's going to throw out my game.

Many thanks

-----

This is the robot in question:

_________________
Blade: Ross Leidy custom blade
Forehand: Butterfly Tenergy 05-FX 1.7mm
Backhand: Dr Neubauer Gangster and Desperado 2 (testing both) OX

Butterfly Amicus Professional robot
-----
Ever tried, Ever failed? No Matter.
Try again, Fail Again, Fail Better!

Chris,
The quality of robot makes a big difference. We have something similar to that model at work and whilst provides a good workout, the spin/speed ratio is unrealistic. In contrast, the robot they have at BATTS is really good. I don't know the model but it cost around 7-800 gbp I believe. I guess you get what you pay for.

There is also a robot at northlands you could take a look at: as a team member if we have a free week we can use the facilities.

_________________
Timo Boll ALC ST
FH Tibhar Evolution MX-P Max
BH Tibhar Evolution FX-S 1.8
185g

so_devo is right - the quality of the ball delivery makes a big difference. Depending on what you're trying to achieve, of course. I had a 989H for a while, but the ball delivery was a low, flat arc. Made short backspin shots very difficult to set up, and the topspin was way too low trajectory to be realistic. It was still good for footwork exercises, blocking, away from the table looping, etc. The programming of this robot was really, really good.

I replaced it with an old Amicus 3000+, and the ball delivery was much more natural. Ball release made a far more normal arc, speed and spin could be easily setup to be very much like a real game. The landing point of the ball was not as accurate as the 989H, but I quite like that. Keeps you on your toes. The Amicus was a much better machine for me and my needs.

yes it is hard for a robot to brush loop, so the spin isn't perfect
I would adjust the head angle and try and find a ratio that will give you the best result
Having said that the benifits of a robot are
footwork excerises, ie falconberg drill etc
serving practice,
playing time
and sometime you can set it on a hard to retrieve serve lol

_________________
Blade Ulmo Duality| FH Tibhar mx-p Black, Dawei 388D-1 red OX
NZ table tennis selector, ask a question
My Blog here..How table tennis objects are made
Table Tennis abbreviations, and acronyms

I have the same observation with my newgy 2050 robot. After practicing with it, it takes me 3 games with an opponent to adjust to the spin on the real world. But the advantage keeps me using my robot. I gained on foot speed and improved my loop. And the games seems to be slower.

BTW chris, your robot looks the same with a newgy 2050.

_________________
Enjoy the game. It is a never ending learning experience once you play and keep the ball on that 45 sq ft of space or should I say half of that space.

Thanks for the replies everyone

As a number of you on here have picked up, it's the spin / speed ratio that is causing me problems. The balls produced do not bounce like a ball I'd get from an opponent as they bounce quite flat with a low flat arc.

At least this means that when I do buy myself my robot (which will be a top quality one), I shouldn't experience these issues.

_________________
Blade: Ross Leidy custom blade
Forehand: Butterfly Tenergy 05-FX 1.7mm
Backhand: Dr Neubauer Gangster and Desperado 2 (testing both) OX

Butterfly Amicus Professional robot
-----
Ever tried, Ever failed? No Matter.
Try again, Fail Again, Fail Better!

Good luck on your purchase Chris.. Enjoy!

_________________
Enjoy the game. It is a never ending learning experience once you play and keep the ball on that 45 sq ft of space or should I say half of that space.

Chris, I have a similar model of the SuperMaster for testing here at Newgy with two wheels. I too, dislike the spin/speed ratio on the balls it delivers. Making only one wheel spin forward while setting the other at 0 does not give good speed and the shot is unstable. As soon as I make the opposite wheel turn, the spin on the ball becomes very low, even at the speed of 1. And as I turn the speed up even more on the second wheeI, I just get an ever faster, even deader ball--something I rarely encounter in a real game. I find it very difficult to adjust the wheel speeds to get balls that I like to practice against, both slow and fast.

I find this same general behavior with almost all 2-wheel robots that allow the wheels to only spin forward, including Y&T, KillerSpin, AMDT, and Oukei. I find the lack of spin produced by these robots to be disappointing. In essence, the second wheel is used only to detract spin, which might be good if you want to practice against dead balls, but really is limiting once you play with players that like to spin the ball hard.

I also found this to be true with the 3-wheel Amicus Pro (first version--I think they're now referring to this as the Amicus Basic). The spin/speed ratio is much lower than I typically get against trained topspin players. I'm hoping to get the new Amicus Professional control box and see if they have improved on this aspect of that robot.

For my own practice, I prefer the 1-wheel Newgy 2050 (disclosure--I helped design this robot). I find the spin/speed ratio to be more similar to shots that I typically get from a trained topspin player with inverted rubber. With one wheel, spin increases proportionally with the speed of the wheel which works out well for 85-90% of the shots typically encountered in the usual topspin rally--counters have moderate speed/spin and loops carry considerable speed plus spin. These one-wheel robots cannot do no-spin returns or slow heavy spin loops or spinny floating chops, but these are not as important, IMO, as getting a good speed/spin ratio on shots typically encountered in a fast topspin rally.

Please also realize that all robots, including Newgy, do not hit the ball like a human and therefore it will feel slightly different when playing against human-produced shots. 1-wheeled robots are more similar to how a human contacts a ball (one contact point for applying speed and spin) than 2-wheeled robots (contact on two different points). But neither 1 or 2 wheel robots will have the exact same feel as a shot from a human.

Robots are a good training tool, but they need to be incorporated into a more complete training program that includes coaching, one-on-one practice, practice games (where the object is to work new skills into a game rather than to win the game), multi-ball, competition, nutrition, physical training, mental imagery, videotape analysis, etc.. You may want to read Chapter 4 of the Newgy Robot Training Manual (http://www.newgy.com/docs/cms/instmanua ... Manual.pdf) for more discussion on this subject of how to transition from robot practice to human practice.

Good luck.

Larry

The SuperMaster is not even close to being like a Newgy 2050 in my admittedly biased opinion. SM is a Chinese Newgy-knockoff with two wheels. Newgy has one wheel and with much higher build quality, IMO.

The Control Box is vastly different--the SM has basic controls for wheels speeds, sideways head movement speed, and frequency (all which cannot be changed during a rally). The Newgy 2050 has these basic controls, but with a lot more options--variable wheel speeds and ball frequencies during a rally, much more accurate and changeable ball locations, ball counter and timer, 64 pre-programmed drills, left-right switch (so asymmetrical drills run correctly for right- or left-handed players, or alternatively, provides a simple way to have a new drill variation), calibration controls (so one robot performs similarly to another), and you can connect the 2050 to a PC to modify the existing drills, create new ones, use ones sent by a friend/coach, or run the robot direct from your computer.

You can even design drills with very advanced programming like a variable number of shots to one location before switching to another location (1, 2, or 3 shots to the BH before throwing a ball to the FH), or throwing to one location and then randomly choosing a location for the next shot (first ball to the centerline, second shot to either the wide FH or wide BH). These are things that even robots 4-5 times the cost of the 2050 can not do.

The 2050 also has a much wider oscillation range, more reliable ball feed, greater frequency range (up to 170 balls per minute!), better build quality, and industry leading customer service, repair, and documentation. SM requires you to self repair or return your robot to China and their documentation is sparse and in badly translated English.

And if you don't want to take my admittedly biased opinion to heart, here's an unbiased review by Greg Letts of Australia, who used to be the TT Guide on the About.com TT forum:

http://tabletennis.about.com/od/tablete ... Review.htm

Larry

I just bought a robo-pong 1040, and I actually really like it. I set it up a little behind the table on top of its shipping box, and it seems to make the trajectory more realistic (I actually got the idea from someone on this forum, don't remember who though). Spin seems good to me, especially when I'm trying to topspin a backspin shot. I feel like it's good for beginners and novices like myself, who need to hone their basic strokes. I hardly ever use the oscillation feature myself, at the moment I'm really trying to fix my UGLY stroke, but when I do it is definitely a workout. I'll usually just hit forehands for a bucket and just backhands for another, then alternate. I don't have a net, but if you get it through Amazon you get a ball picker up thing with it, it works pretty well. I've read reviews complaining about the balls getting stuck, but as long as you keep them clean it's usually fine (and I have the cleanest balls around). I've got about 100 balls, and by the end of a bucket my legs are killing me, so it's a pretty good workout that way too. Anyway, I guess the bottom line is how you plan to use it and how much you want to spend. Good for practicing strokes, probably not so good for simulating real gameplay, but you have to spend a mint to get one that does that, and I don't know how accurate that would be anyway. Good luck, hope you find one you like!

I purchased the Y&T V-981 Robot and it's a decent basic robot. I mostly use the top spin to practice my forehand and looping. The underspin is unrealistic (or I haven't figured out the right combination to make it realistic) but it does a good enough job for me. My whole setup with the ball collection net and balls was about $400US.

This is where I got it on ebay: http://www.ebay.com/itm/Reliable-basic- ... 3388cc6b88

Just trying to get this straight. Wouldn't a one wheel robot be the equivalent of a two wheel robot with one wheel not spinning?

Iskandar

Probably not, as the wheels have high friction so they can generate spin. If the robot has only one wheel other side will probably have less friction.

In my club we occasionally use a Newgy Robopong 2040 with only one wheel. The Topspin strokes I think are quite ok, the placement of the balls is very bad as the head as a lot of play. Backspin is unrealistic high.

Kind Regards

It would seem that way, but most of the two-wheel robots I've tested have less stable trajectories than one wheel robots, even from the same company. I think the problem is that the wheel that supposedly is being held still does actually move as the ball is squeezed between the two wheels and pushed out. So the stationary wheel could move a little, a lot, or in between; each time producing a slightly different trajectory.

With a stationary rubber block (or other similar structure) to force the ball into a single spinning wheel, the ball seems to be held in a much more consistent way as it is pushed into the wheel and is squeezed between the wheel and the rubber block.

Something else I've noted along these lines with the two-wheel robots is that as soon as both wheels spin forward (the vast majority of two wheel robots are designed to only spin forward), the ball becomes quite dead and a dead ball being thrown through the air is inherently less stable than a spinning ball. I noticed this on the Super Emperor I tested--at top and bottom wheels set at 1, there was probably a 2 to 2.5 foot difference between the landing spot of the thrown balls. One ball would land on the table about 1-1.5 feet in front of the endline and the next ball would be thrown off the end of the table by 1-1.5 feet.

Larry

The looseness of the head on the 40 series Robo-Pongs is largely determined by the settings of the Oscillator levers at the rear. Setting these at 3 and 4 should result in the head being held pretty tightly and the side to side location should not vary too much for a static delivery. If the levers are in positions other than 3 & 4, then there will be more looseness in the head if you were to stop the head, but this looseness wouldn't affect anything as the head is moving. The loosest setting would be with the levers set at 1 and 6 which results in no oscillation and is intended only to protect the oscillation mechanism from damage during transport.

The 50 series Robo-Pongs have a completely different oscillation mechanism in which the head is held at the same tension no matter where it is. There are no levers to set as the side to side location is determined solely by the servo that drives the head horizontally.

Regarding the backspin, you should be able to set the backspin high or low. If it is low, it will probably be a bit faster to land at the same relative location on the table as a slower backspin shot. The head angle adjustment is what determines the height of the shot and it can vary from just barely skimming above the net to about 2 or more feet over the net.

Larry