Skinner-box rats trained to predict currency market movements

What about mice? :stuck_out_tongue:


Funny - this is exactly the approach used in machine learning, only using a really inappropriate tool for the task: rats rather than statistical models. Mind you, the investment guys are probably tools too…


And rats seem to have morals…


Rats are more interactive. Rats see you as a big rat - mice see you as furniture.


Hmm, my mouse (Mousey) doesn’t see me as a furniture and is interactive. :stuck_out_tongue:


Yes, a lot of people seem to be missing the point. This was created by an artist, not a scientist—presumably to point out how insane it is to offer huge compensation packages to Wall Street guys who happened to get a lucky run here or there. Because it’s a statistical guarantee that some guys on Wall Street are going to have a lucky run here or there regardless of whether they know what the hell they’re doing or not.

I see this as a clever way of saying: Hey, Mr. Wall Street—prove you deserve that big bonus more than my rat.


Either it’s a joke, or this guy’s an idiot. Just doing a quick check on the full litters of his third generation rats, 50% is well within the standard error of the mean for both litters, taken either separately or together, meaning that there’s no evidence presented that the breeding had any effect.

That said, my money’s on joke, since it provides a nice commentary on the fact that real traders have pretty similar spreads to this, and any time you see a real trader consistently beat the market (except through insider trading) it’s not by much and not inconsistent with it just being random chance based on the pool of traders out there.

A negative result is also a successful experiment. If the hypothesis can be either proved or disproved, the experiment is successful. Too bad data for either outcome counts as failure.

Except that he says: “The results were astounding: the second generation of top traders had a much better performance than their parents, but undoubtably, further research will be necessary to confirm this finding of ours.”

Indicating that he hasn’t the faintest idea what he’s talking about, or it’s a joke.

The rat is a bio-assay for a neural network.

They’re experimenting on us!

I don’t believe this. It would likely be illegal under EU laboratory animal regulations, and nobody has the space and time to breed new strains of lab rats privately so it would have to be done in a lab so it wouldn’t get approved.

Are we talking the cocaine and hookers study group, or the controls?

1 Like

The EU may have their consistency a bit more in order; but at least in the US it’s amazing what you can get away with as long as you aren’t doing science. If your purpose is to pet it, eat it, or make a coat out of it, you enjoy broad latitude. If you are planning on learning something, though, better put on your mercy hat and prepare to face the IRB…

As mentioned above, he is an artist, not a scientist. It’s probably better to regard this as a piece of art rather than a joke or serious science.


Ahh. I am corrected.

A LOT of things can be done privately. There are (admittedly rare but existing!) people with the resources, and it’s amazing what you can do for pennies on a dollar with just a bit of creativity and tech skills.

There is even an emerging sector of DIY science, with real (and usually improvised) high-end lab equipment. There are even budding vendors making money on it, and more (including me) aspiring for that.

There are more and more regulations on “big” institutional labs, more commissions, and one has to do a week of paperwork for a day of research. A little competition from the garage-lab sector is only a good thing.

1 Like

I used to have domestic rats as pests, they were horrible. I hate rats!

1 Like

Teela Brown is unimpressed, but doesn’t quite get what’s going on here.