The Randomizer

Tuesday, May 19, 2009

Another day, another trip through the grist-mill of seemingly random (and certainly erratic) price action. In some ways, it doesn't seem to matter if you are bullish or bearish; if you're trading tactically, it's very, very easy to get caught out by the seeming randomness of it all.

So easy, in fact, that Macro Man wonders if it is even possible to tell the difference between current price action and a random number generator. So he's determined to find out.

There are six charts below. Three are actual financial market prices, and three are the product of the random number generator function in excel. Your mission, should you choose to accept it, is to determine which three are real....and which three are the figments of Bill Gates' imagination.

Good luck.

Posted by Macro Man at 9:35 AM  


F, E, B

Anonymous said...
10:25 AM  

I suppose I should be able to guess some of the others from your favourite assets from previous posts, but haven't quite got round to looking yet...

Anonymous said...
10:37 AM  

A, C , F its excel

Anonymous said...
10:40 AM  

i think A is USD/JPY, B = EUR/USD and C = Stock of some description.

Anonymous said...
10:50 AM  


t said...
11:15 AM  

c, d and f

Anonymous said...
11:53 AM  

B, C, F

DVP said...
1:12 PM  

maybe they are all random and MM is conducting a social experiment?

Anonymous said...
1:31 PM  

A,D,F are markets.

. said...
1:39 PM  

Mmmm, tough one.

However, chart A is definately Yen pr dollar, and chart E looks like CAN$ pr Swiss Franc to me.

For the rest, I don't know, excpet that B is NOT €/$.

Tord Steiro said...
1:51 PM  

I'm baffled. But wouldn't it be challenging (maybe not impossible, but at least challenging) to tell the difference in more "normal" times?

Bruce Chadwick said...
1:55 PM  

It was in the mid 1960's when mathematical economists began studying this phenomenon. Paul Samuelson wrote a paper that attracted wide attention - "Proof that Properly Anticipated Prices Fluctuate Randomly". Thought you'd be interested. Here's a copy on the web:



tas said...
2:28 PM  

B, D & E are is excell

InEgoVeritas said...
2:47 PM  

F looks like a buy here

Anonymous said...
3:00 PM  

A,B & F are market prices. Fred.

Anonymous said...
3:16 PM  

Isn't there an online test that does this? I'd search, but I have to run to a meeting.

Also: do not use Excel's generators for anything important, especially if you care about the tails. They are still, last I heard, broken.

Real men use R.

wcw said...
4:58 PM  

The deteriorating quality of your posts can only be explained by the losses you are taking in your trades as well as investment views.
What kind of nonsense is this? I tried to put up with your stuff on renaming assets (breaks anchoring, so a good approach). But this is just plain stupid!
Take a break. You blog well.

Anonymous said...
5:26 PM  

Grrr, it's not the usd/jpy. It could be an obscure currency couple or MM rolled the voodoo bones and plotted the numbers.

I have no time today and this puzzle will bother me considerably.

Is it the number of portfolio managers leaving London to prevent the seizure of their wealth?

Professional Gringo said...
5:40 PM  

Anon @ 5.26, the point is that if you can't tell the difference between what you're trying to trade and the output of a random number generator, perhaps you need to take a step indeed I have done.

Given that no one has gotten the right answer, perhaps it's not as stupid a lesson as you might think.

Macro Man said...
5:52 PM  

no one has gotten the right answer?? Damn. Well at least there is a chance I was right then. B, C, and D look random. But I'm not at all sure of that, unlike the last time I saw this kind of comparison.

shan said...
8:02 PM  

Well, what you are implying is that the current price action is somehow very random and somehow that is not the case in 'normal' markets.
Give me example of one time series for any asset such that you can pick it out of a list of random excel charts as being the price action of a real asset.
Ofcourse they all look random. And the purist in me is kinda glad because of that.
To trade is to stop whining and fighting the tape.

Anon@5.26 said...
9:46 PM  

A) Much of the price action of the past two years has been more clearly non-random than the past two months. That change in market tone is worth pointing out.

B) If what you read on a free website written in his spare time by a guy forced to work from home for most of the last three months doesn't meet your lofty standards, consider your objections noted and feel free to spend your time in more high-quality, non-whiny locales.

Macro Man said...
9:54 PM  

I believe that Anon 5.26 meant to say "the Macro Man blog has a certain je ne sais quoi and gives me big morning wood."

As for the puzzle, look at the clues. Price on the left and date at the bottom. The charts have different prices but the dates are the same. The patterns could be anything.

Are you sure that this can be solved with that information?

Professional Gringo said...
11:03 PM  

B C and F are real as they are going up.


Anonymous said...
11:13 PM  

i've been playing with this approach recently (using R that wcw suggests -- free ware -- fits my price range, but, I think, very good, used in stats grad programs at many places)

i've been literally creating random walks (adding normal -- mean zero, and some standard deviation to starting price).

it's not hard to get things that look a lot like a general financial time series -- even things that look a specific time series

its likely that any statistician worth his or her salt can choose functional forms that come very close to anything you want (with the caveat being that you can do a great job of modeling what happened IN THE PAST -- and if you believe in your model too much when the regime changes -- as MM suggests may be happening -- you find all your fancy stats falls apart -- e.g. Renaissance Technologies not quite the superstar today that it was in the past -- fell off Barron's list of top hedges funds based on last 3 years)

triozyg said...
12:30 AM  

just for fun i will take a stab at it (I mean, jeez, there are only 20 possible answers – 6 choose 3)
A looks like usd/jpy (re: 10:37)
B looks like eur/usd (re: 10:50, disagreeing with 1:51)

since I think A and B are right and MM ruled out ABC and ABF (since they were proposed before MM said all were wrong at 5:52) then I'm going to choose between ABD and ABE, going with the idea that C and F are examples of excel's random number generator.

To my eyes C and F have odd “curved” segments to them that A and B do not have, so looking at D and E it seems to me that D has one of those odd curves in it, so I’m going with E – and anyway, I like ABE, one of my favorite presidents.

Of course, I have no clue what E is

triozyg said...
12:41 AM  

I'm going with a,d,f

working on the theory that positive serial correlation is more likely in a real market than a random number generator

Awake said...
1:21 AM  

Ok, cool.

Now, can we have a second chart, charting the choice distribution by the people commenting?

I think that might tell us something interesting....

CTMM said...
1:23 AM  

ACF are the real markets. Nothing is random.

Anonymous said...
3:44 AM  

B C and D are random. No science, just hunches on the basis of looking at one or two charts in the past.

Get well Macro, you will need all your strength to battle the PPP without the SEC. You can always fall back on reworking Houseman if the Tiny Tim ends up owning every share on every exchange. Funny, saw it on my dad's shelf last I saw him. Never knew that was his style.

H(oratio) said...
6:19 AM  

F mkt

If B is not euro, then MM's excel is very fishy.

Abhishek said...
6:44 AM  

So cute! Thank you!
Great template! Thanx for sharing it with us..... more templates

rasitha said...
10:32 AM  

I must admit that B looks a lot like dollars pr euro until the 7th of May. After the 7th, the graphs do indeed look different, which is why I stated so clearly that it could not be euro pr dollars (or the inverse, of course, which it looks like).

The third one I still don't know, and my suggestion for the E chart doesn't seem to fit considering the entire period.

Further, given the hints, perhaps the point is that the current market IS random, so it doesn't matter if it's generated from excel or the some kind of market.

Tord Steiro said...
12:06 PM  

Post a Comment