Market Models and Randomness

[SIUYA 11]

An interesting article on markets and movements from a thought experiment and resoning point of view for those interested.

 

http://www.bloomberg.com/news/2012-02-07/a-bar-may-be-best-place-to-understand-markets-commentary-by-mark-buchanan.html

 

and the link in the article to about inductive reasoning and how it can apply to markets.....and forum threads/posts

 

http://phys.ubbcluj.ro/~zneda/edu/mc/minority-game2.pdf

[bobcollett 10]

An interesting article on markets and movements from a thought experiment and resoning point of view for those interested.

 

A Bar May Be the Place to Understand Markets: Buchanan - Bloomberg

 

and the link in the article to about inductive reasoning and how it can apply to markets.....and forum threads/posts

 

http://phys.ubbcluj.ro/~zneda/edu/mc/minority-game2.pdf

 

I would like to reopen this thread.I think it got lost in the ether.

It is a bit hard to follow, but its offers some eye opening ideas on deduction

Especially the ability to see patterns more easily than pure analysis.

Thus the big TA following with charts

And all those posters on "Dont be fooled by randomness" should have a look.

regards

bobc

[SIUYA 11]

I dont follow the maths too much but there is a lot more recent information on the mathematics of tipping points that probably is similarly related.

 

There was also an article that uses riot mathematics to talk about how crowds either get going into a riot or dissipate quietly. Tim Harford touches on such subjects that can lead to further investigation if you like.

 

Tim Harford — Article — The random side of riots

 

basically while most might be rational, they have a tendency to become irrational not because their mind set has changed but because others around them have....

[BlueHorseshoe 10]

The issue that I have with this analogy is that several of the inputs are more static than they would be in most market scenarios. The club's capacity remains constant from week to week, as does the population of potential attendees (or "clubgoers", as I believe they're called ).

 

Traded price can be viewed as a noisy measurement of an asset's true underlying value, but this value is not constant and changes over time according to a non-linear process.

 

My guess is that rather than trying to incorporate fat tail events - sudden non-stochastic price jumps - into the model, it would be better to use them as a cue to reset the model. The model, after all, is not built to measure them. Maybe this explains why so many HFTs supposedly step aside when volatility rises?

 

BlueHorseshoe

[wmck6167 10]

"http://www.bloomberg.com/news/2012-0...-buchanan.html....."

 

I noticed a comment:

 

"Steeloak 11 months ago 1 comment collapsed Collapse Expand

As a trader, I long ago learned not to guess why the market is doing something, I simply wait to see what the market is doing & get onboard the trend. I do trade around news events, but only to see how the market reacts, not to guess what it will do.

 

Other than that, it is simply cut my losses short & let my winners run. Sounds easy, but it has taken me 5 years to get good at it.

 

I think anyone who says they can predict markets, or knows the reasons why the market is doing what it is doing is just talking through their hat."

 

I just wonder why he handled simply cut losses short and let winners run took so long to get good at it..... Indeed this is the question I am thinking to about it this year. I think days and nights, and even in my dream..........

 

Anyhow, thanks for bring out this question and have a good profit trades next week.

[SIUYA 11]

The issue that I have with this analogy is that several of the inputs are more static than they would be in most market scenarios. The club's capacity remains constant from week to week, as does the population of potential attendees (or "clubgoers", as I believe they're called ).

 

Traded price can be viewed as a noisy measurement of an asset's true underlying value, but this value is not constant and changes over time according to a non-linear process.

 

My guess is that rather than trying to incorporate fat tail events - sudden non-stochastic price jumps - into the model, it would be better to use them as a cue to reset the model. The model, after all, is not built to measure them. Maybe this explains why so many HFTs supposedly step aside when volatility rises?

 

BlueHorseshoe

 

good point.

 

Comparing the market to riots....

 

Maybe given the huge number of participants (hence the average might be similar), the common goal of making money and the huge influence of index trackers etc, maybe the population of the markets and the inputs is itself relatively static (apart from a few excessive times of volatility).

As for the changing values of an asset, are not the morals/values/concerns of society also changing over time.

Also isnt this one of the main points of the randomness arguments - be it crowds or markets. Why is it that normal event s can be triggered into abnormal events. For markets drifting away from value is normal, but being able to consistently proift from it should not be.

[BlueHorseshoe 10]

good point.

 

Comparing the market to riots....

 

Maybe given the huge number of participants (hence the average might be similar), the common goal of making money and the huge influence of index trackers etc, maybe the population of the markets and the inputs is itself relatively static (apart from a few excessive times of volatility).

As for the changing values of an asset, are not the morals/values/concerns of society also changing over time.

Also isnt this one of the main points of the randomness arguments - be it crowds or markets. Why is it that normal event s can be triggered into abnormal events. For markets drifting away from value is normal, but being able to consistently proift from it should not be.

 

Yes, the flip-side to complaining about aberrant models is probably to acknowledge that to make money one only needs a model that works more often than not (or rather, loses less when it's wrong than it makes when it's right).

 

Being pedantic but . . .

 

the common goal of making money

 

As you have pointed out yourself in other threads, the ways in which many participants make money has nothing to do with whether they're buying or selling the particular market you trade right now. StatArb, actual arbitrage, hedging, re-balancing, gaming another participant, or just executing a very large order for a rich but naive heir - there are countless reasons to go to the club, even if you don't like crowds or dancing . . .

 

I find it interesting (I've been reading Taleb's 'The Black Swan' recently) that even those who acknowledge randomness don't really know how to profit from it in addition to profiting the rest of the time. Probably it just isn't possible.

 

BlueHorseshoe