Robot Scrum Machine

[Groucho]

I think that's where super-computer complexity that Bruce refers to comes in.

Nah, the software simulation is the easy part, a straightforward problem in finite element analysis. I could write the code.

I've written code for flight simulators, in my distant past.

 

[Groucho]

Just as importantly, the opposing front-rower is continuously reacting to the forces exerted on him to gain an edge - could a robot replicate that element?

Yes.

 

[Bruce Ross]

I used to use the ScrumTruk a bit in my days at Uni. I was never game to try anything much over 300kg as a front rower. A back hefting 400kgs? No wonder Chariots is so lightning fast and good looking! But seriously it is a bloody good machine and a secret weapon in the Uni arsenal.

Thanks, MrMouse. 300kg on the ScrumTruk is serious tin. The problem with a lot of good equipment is that people are not prepared to explore their limits. As you would know, simply working up a sweat in a gym achieves nothing.

I am working my arse off to try to stop it being a secret weapon. I need to make money out of these machines.

 

[Bruce Ross]

Bruce, leaving aside the "Packing down against England on a wet surface with Sheridan in a bad mood" options, from an engineering perspective what is being suggested bears a close resemblance to the problems addressed by flight simulators, or advanced driving simulators, and probably presents a similar level of difficulty.

No, Groucho, let's not leave that aside, because it is precisely the problem that was posed by doug. With this suggestion you remind me of economists - I used to be one - who assume away all the complexities to produce a solution reducible to a two dimensional diagram.

And so you postulate "the front row as a single plane. A more sophisticated version would consist of three such planes." And of what possible usefulness would such a simple contrivance be? Doug was wanting a simulator built that would be "our secret weapon for the World Cup".

I presume you haven't spent much time inside the middle of a scrum.

But whilst a front-rower may be exerting all his influence through his shoulders and bind, he is pivoting not only through the line of his shoulders, but also his spine, hips and knees - I think that's where super-computer complexity that Bruce refers to comes in. Just as importantly, the opposing front-rower is continuously reacting to the forces exerted on him to gain an edge - could a robot replicate that element? Of course, in Kiwi mode, the robot would need an extra piston to replicate the Myth's hand on the deck tactic.

These are just some of the complexities before factoring in the forces generated by the players in the middle and back rows, the strength profiles of each of the players, their degree of fatigue, etc., etc.

 

[dougdew99]

A starting point might be a machine which measured the force and direction being exerted by the packed down scrum... this might enable continuous refinement and improvement... show the effect of meerkatting... enable comparison of the effectiveness of individuals, etc., or do we already have this?

 

[Groucho]

No, Groucho, let's not leave that aside, because it is precisely the problem that was posed by doug. With this suggestion you remind me of economists - I used to be one - who assume away all the complexities to produce a solution reducible to a two dimensional diagram.

And so you postulate "the front row as a single plane. A more sophisticated version would consist of three such planes." And of what possible usefulness would such a simple contrivance be? Doug was wanting a simulator built that would be "our secret weapon for the World Cup".

I presume you haven't spent much time inside the middle of a scrum.



These are just some of the complexities before factoring in the forces generated by the players in the middle and back rows, the strength profiles of each of the players, their degree of fatigue, etc., etc.

Gee Ross, don't get me wrong please. I'm not being critical. I'm simply interested in the engineering question.

No, I haven't spent any time in a scrum. However I've been a member of a team that built a full 3D flight simulator apparatus, for missile simulation, for the RAN. I think in engineering terms this in an intermediate problem in robotics, which is why I suggested you speak to an engineer at Sydney University if you were interested in the idea. When we built the simulator for Harpoon, we used as the starting point a rig developed by students at MIT.

I fully understand you can't model all of the forces present in 16 bodies, but that doesn't mean a device cannot be useful.

But then again, I'm a physicist by training, not an engineer. I shall leave questions of feasibility to practitioners.

 

[Jnor]

No, Groucho, let's not leave that aside, because it is precisely the problem that was posed by doug. With this suggestion you remind me of economists - I used to be one - who assume away all the complexities to produce a solution reducible to a two dimensional diagram.

Didn't know your alias was Frank Stilwell, Bruce?!?

Anyway isn't this all a bit silly? The only thing that'l get us up to a better srummaging (and general forward play level) is focussing on it better at all levels from grassroots to elite, as has already been said. Opposed scrums with good opposition, not Terminator-TPN hybrids, as desirable as that would be...

 

[Bruce Ross]

Gee Ross, don't get me wrong please. I'm not being critical. I'm simply interested in the engineering question.

No, I haven't spent any time in a scrum. However I've been a member of a team that built a full 3D flight simulator apparatus, for missile simulation, for the RAN. I think in engineering terms this in an intermediate problem in robotics, which is why I suggested you speak to an engineer at Sydney University if you were interested in the idea. When we built the simulator for Harpoon, we used as the starting point a rig developed by students at MIT.

I fully understand you can't model all of the forces present in 16 bodies, but that doesn't mean a device cannot be useful.

But then again, I'm a physicist by training, not an engineer. I shall leave questions of feasibility to practitioners.

I think we are getting close to finding common ground.

And, Groucho, my genuine and heart felt apology for comparing you to an economist. Physicists do have very high range analytical skills.

 

[Groucho]

I think we are getting close to finding common ground.

And, Groucho, my genuine and heart felt apology for comparing you to an economist. Physicists do have very high range analytical skills.

If I understand forward play correctly, a device would need a hand to reach and squeeze the tighthead's balls.

 

[Bruce Ross]

Didn't know your alias was Frank Stilwell, Bruce?!?

For all his criticism of neoclassical economics, Jnor, to my knowledge Frank is a believer in economic theory. I on the other hand regard it as a crock.

 

[Bruce Ross]

If I understand forward play correctly, a device would need a hand to reach and squeeze the tighthead's balls.

In the modern scrum, Groucho, that hand would have to come from his own side of the scrum. Similar tactics are not unknown as a means of getting a dumb prop fired up.

 

[Jnor]

All economic theory? If it wasn't so late now there would be a discussion to start!

I'd be interested in what brought about the change of heart as well...

 

[MrMouse]

Opposed scrums with good opposition, not Terminator-TPN hybrids, as desirable as that would be...

So my TPN-Xman hybrid is still on?

As a squad exercise, it would be interesting to get the Wobs boys at Victoria Barracks scrummaging down against a pack of the best scrummagers in club rugby. They don't have to be supermen around the paddock, potplants welcome, just scrummage the shit out of em a few hours a week. Pay them a stipend if necessary, it seems fair. We don't need superhumans or Andrew Sheridan in our back yard for that.

 

[MrMouse]

Thanks, MrMouse. 300kg on the ScrumTruk is serious tin. The problem with a lot of good equipment is that people are not prepared to explore their limits. As you would know, simply working up a sweat in a gym achieves nothing.

I am working my arse off to try to stop it being a secret weapon. I need to make money out of these machines.

At that weight keeping form was a serious challenge and if I went in tomorrow I'd have a heart attack by 200 probably. I think the occasional jog around a paddock would have explored my limits

You seem to be developing uses in a range of sports - you just need greater coverage within sports/sports media. Or you need to find a benefit to sell to commercial gyms and get it rolling out the door in a different direction. Highly adjustable to any user, core stabilisation, good for those with back problems etc etc all seem like winners to me. Any chance it could have a role in rehab? Good luck with it!

 

[Crow]

I presume you haven't spent much time inside the middle of a scrum.

These are just some of the complexities before factoring in the forces generated by the players in the middle and back rows, the strength profiles of each of the players, their degree of fatigue, etc., etc.

Those complexities aren't as difficult to deal with as you believe. People have been modelling full human motion in realtime for the last 5 years.

I've worked with haptics in the past and currently with high performance computers. If you were willing to spend the money to build the device, the kinematics computations themselves are well understood and reasonably easy to model. I would go far as to say that a reasonably powerful desktop machine could handle the computations easily (there are limits, of course.). A HPC would be overkill.
I don't know of any haptics devices large enough to do this type of thing, though. So that would need to be developed.

The problem would be in telling the machine what to model and why. I think that would require significant time and experimentation, which eats the money very quickly.

 

[Rob42]

So Bruce, for someone who's never trained with big steel machines, what does Tom Carter's 400kg mean? I mean, I can push a car weighing more than 400 kg horizontally, but this machine's obviously providing more resistance than that.

 

[Groucho]

Those complexities aren't as difficult to deal with as you believe. People have been modelling full human motion in realtime for the last 5 years.

I've worked with haptics in the past and currently with high performance computers. If you were willing to spend the money to build the device, the kinematics computations themselves are well understood and reasonably easy to model. I would go far as to say that a reasonably powerful desktop machine could handle the computations easily (there are limits, of course.). A HPC would be overkill.
I don't know of any haptics devices large enough to do this type of thing, though. So that would need to be developed.

The problem would be in telling the machine what to model and why. I think that would require significant time and experimentation, which eats the money very quickly.

Crow, this seems quite analogous to a wrestling problem (or three wrestling problems).

I've had a look around and I can't find a fully-featured wrestling machine, but there is a plethora of arm wrestling machines. These are interesting because they provide a fixed base implementation of human arm movement, and are therefore probably closer a scrummaging problem that a fully-featured wrestling machine would be, since that would have to move in three-dimensional space.

There are plenty of robots that wrestle other robots, but this is obviously a much simpler problem.

Interestingly, arm wrestling machines are fertile ground for academic research into haptics devices. A quick Google yields lots of links.

The kinematics are straightforward, as you say, but the game playing strategies perhaps not so. But it is conceivable that a machine learning AI (which can almost be had off the shelf) could produce new techniques, in the same way that chess AIs have.

 

[Bruce Ross]

So Bruce, for someone who's never trained with big steel machines, what does Tom Carter's 400kg mean? I mean, I can push a car weighing more than 400 kg horizontally, but this machine's obviously providing more resistance than that.

My view is that it would compare to a very heavy squat. As an example, like other Sydney Uni players, Jerry Yanuyanutawa used the ScrumTruk as a regular part of his training. Jerry was prodigiously strong as a squatter, box-squatting 300kg in 2008. But as far as I know he never used anything like 400kg on the ScrumTruk.

Tommy is one of the few who attempted maximum effort on the machine. Matt Williams told me that the Ma'afu brothers, Salesi and Campese, were doing mid 300s when West Harbour had a ScrumTruk, but I don't know whether they were using it from a 90-90 flexed position which is the recommended start position.

The West Harbour machine is now at Newington who seem to be making good use of it.

I think Tommy's 400 effort would certainly equate to a 200kg+ squat set.

 

[Bruce Ross]

You seem to be developing uses in a range of sports - you just need greater coverage within sports/sports media. Or you need to find a benefit to sell to commercial gyms and get it rolling out the door in a different direction. Highly adjustable to any user, core stabilisation, good for those with back problems etc etc all seem like winners to me. Any chance it could have a role in rehab? Good luck with it!

Thanks for the endorsement, MrMouse. We are gradually rolling out a range of machines and the commercial gym and rehab markets are very much in our future plans.

Since June we have had an English company manufacturing and distributing MyoQuip machines throughout Europe. One of their initial sales has been to the Exeter Chiefs who have just been promoted to the Premier Division. They installed the MyoTruk and the MyoThrusta. They have already collected some impressive scalps.

They also sold the same machines to the London Rowing Club where Sydney Uni's ex-Director of Rowing Phil Bourguignon now coaches. He is already reporting 8-10 sec improvement in the erg scores of his rowers, presumably a combination of Aussie coaching and Aussie developed equipment.

 

[Boomer]

Holy crap - the French have done it...

While the Wallabies are preparing for their daunting battle against France in Paris on Saturday by training against a standard scrum machine, their opponents are matching wits with a computer-driven scrum simulator they invented in conjunction with Thales, the company recently awarded a $1.78 billion contract to develop Europe's next generation of weather satellites.

Rage against French machine
http://www.theaustralian.com.au/news/sport/rage-against-french-machine/story-fn6atwd0-1225960476969