Trading System Junkyard

[beginner_av]

defining the above and initial data mining would be a good starting point.

 

[uasish]

Plz Watch how they approach .

 

[oxusmorouz]

What should be the first step in system development?

Defining the instrument. Benchmarking.

 

[raosrinivas]

Defining the instrument. Benchmarking.

i thought Benchmarking is the last thing you do...

 

[oxusmorouz]

Thats good SGM. Lets take that as a starting point. What would you do next?
On the side Sanjay, why and how did you come to that premise (for discussion purposes)?

A = If markets are efficient, follow benchmark.
B = If inefficiency exists, trade trading system signals.

Instead of going into either of these assumptions and testing if these assumptions work on future data set, a simpler solution would be to allow the market to determine which survives, and trade the premise for which the equity curve is higher. If little correlation exists between strategies followed in either assumptions, returns of the trading system will be ~= the higher equity curve and the risk characteristic will be that of the benchmark.

Maybe this answers Srinivas' question too.

 

[raosrinivas]

A = If markets are efficient, follow benchmark.
B = If inefficiency exists, trade trading system signals.

Instead of going into either of these assumptions and testing if these assumptions work on future data set, a simpler solution would be to allow the market to determine which survives, and trade the premise for which the equity curve is higher. If little correlation exists between strategies followed in either assumptions, returns of the trading system will be ~= the higher equity curve and the risk characteristic will be that of the benchmark.

Maybe this answers Srinivas' question too.

oxy, Traders as species exist today just based on the premise that Markets are inefficient and they want to monetize the inefficiencies, doubting market in efficiency is doubting their own existence, if markets were efficient, everyone will invest in an index and find a job. do you think anyone will take the bait for the first ?

 

[oxusmorouz]

oxy, Traders as species exist today just based on the premise that Markets are inefficient and they want to monetize the inefficiencies, doubting market in efficiency is doubting their own existence, if markets were efficient, everyone will invest in an index and find a job. do you think anyone will take the bait for the first ?

If each and every trader had spotted that inefficiency, there would be none going bust. Most traders trade who boast to have spotted some inefficiency would fall under the dreaded "clustering illusion".
Are markets purely efficient or purely inefficient or partly inefficient? If some inefficiency is spotted on training data, would such an inefficiency hold good for future data set as well? Would it continue for an indefinite period of time? If not, when does it cease to exist? If it ceases to exist what effect will it have on my investment returns? If the markets are partly inefficient, when does inefficiency exists and when does it not?
These are questions to be answered if we are to go by a single premise hypothesis (i.e, market is efficient or the market is inefficient). Market "determined" strategy adoption eliminates these assumptions, since we are assured of ~ at least the benchmark returns and at the same time, can trade the inefficiency to ~ its fullest extent. In short, we are just doing away with assumptions necessary otherwise.

 

[uasish]

So we are in no way going to perform below Benchmark .

 

[raosrinivas]

So we are in no way going to perform below Benchmark .

that's what i thought he was pointing to ,

oxy,

i am a man of simple english

http://simple.wikipedia.org/wiki/Wikipedia:Simple_English_Wikipedia

i got confused there.

 

[SGM]

So we are in no way going to perform below Benchmark .

We already outperformed the benchmark , but where's the system