trading stocks: software?

[luxiana]

hello,

is there anybody here using metastock?
it's a product developed by reuters.

how good is it?

[durain]

what you want to know about metastock? what are you trying to find out?

[Thaiclan]

I use CQG and E-Signal...... both great for Algorithmic Futures trading

[luxiana]

hi there,

thanks for your replies.

I was wondering how good was the software:
- is it easy to import csv data (shouldnt be too much of a problem)
- how easy is it to use the technical analysis indicators
- is it possible to monitor a list of stocks with B/S signals dictated by technical analysis indicators?
- Does it calculate EWMA, GARCH volatility for stocks?
- it is possible to work at portfolio level? (both CAPM and Litterman theories, backtesting of a strategy, monte carlo simulations, Sharpe / Sortino ratios)

thanks!

[luxiana]

to be more precise:

I am looking for a software to:
- define a strategy
- monitor stocks
- give me B/S signals according to the strategy
- performs backtestings of the strategy
- analyze the strategy at a portfolio level
- runs Monte Carlo simulations on the stocks and portfolio level
- automate reportings.

[durain]

ok, i will have to ask one of my colleague who is a hard core metastock user!

[luxiana]

hi durain, thanks!

basically I am looking for a software that would help in constructing, analyzing and optimizing an investment portfolio and that wouldnt be limited to just the technical analysis.

+++

[ksl]

to be more precise:

I am looking for a software to:
- define a strategy
- monitor stocks
- give me B/S signals according to the strategy
- performs backtestings of the strategy
- analyze the strategy at a portfolio level
- runs Monte Carlo simulations on the stocks and portfolio level
- automate reportings.

Not possible in my opinion, the permutation of strategies, would be far too complexed, even taking the average, and doing a low discrepency sequence quasi Monte Carlo, would be full or errors, unless you took the worlds best top 20 financial strategists to minimise the errors, and even then, the software would be asking for a response from the users, which again is exponential.

The software programmer could of course input, his own theories of what strategy works best....yet the margin of error is still very large, once a user is asked to act on the results, especially if large amounts of monies are involved.

100 dimensions is by no means unreasonable, since in many physical problems, a "dimension" is equivalent to a degree of freedom and it is like working in space, to ask others to react, because when insight is analysed, more variations of risk aversion instances are involved too, than one is actually aware of.

It would be very interesting to apply all the same, for a person that is very good at risk assessment, and the only way to find out how one rates, is to take a large quantity of test questions from several sources. at least 25 questions X 3 with in my opinion would give a fair rating of ones risk aversion on a scale of 1 to 10...I believe one would have to come in the range of 4,5, or 6 to be called a sound investor, that minimises risk!

[luxiana]

it's all about probabilities:

the simulation at portfolio level would be run when adding one or two components to the basket to check how the portfolio Risk/Return would react in the long run.

[ksl]

it's all about probabilities:

the simulation at portfolio level would be run when adding one or two components to the basket to check how the portfolio Risk/Return would react in the long run.

Okay, thanks I was just looking at crystal ball, to download, although only 15 day trial.

Although I was thinking even with a portfolio of 5 to 10 companies would cause significant errors, due to many factors, that are unpredictable even to input one or two events.

It would be better to isolate and test the probabilities on shares of the same industry, because each company would have different variables, effecting the shares....Personally I treat my stock individually at all times, in the portfolio, it's the only way one can capitalise on the ups and downs, dividends, being another case scenerio, so therefore if you do not act on individual shares, you are not making the money work, and the ROI may end up much less. If you see my point of view.

So you can predict the swings to a certain extent, but if you miss the sell, your ROI will decrease....to maximise ROI, you must swing trade in my opinion, individually.

[luxiana]

CB is pretty cool although I'd recommend @risk.
CB is user friendly but is so slowwwww sometimes.

I've also found @risk tutorials better explained that CB's.

[luxiana]

by the way: what i'm looking for is a "all in one" solution (if possible)

I know I could still use excel + Crystal Ball for MC simulations and various add ons. jsut trying to cut on all the tedious work... (that would involve VBA programming...)

[ksl]

by the way: what i'm looking for is a "all in one" solution (if possible)

I know I could still use excel + Crystal Ball for MC simulations and various add ons. jsut trying to cut on all the tedious work... (that would involve VBA programming...)

Yes I can see that now!

[durain]

hold your horses... i will ask the expert in the office. let me see what we use... bloomberg, t-zero, datastream, thompson, barra, factset, metastock, xtra 3000 (reuter), OMT, instinet, impact, rendezvous, business objects, etc.

[luxiana]

I still remember how to use BBG and Reuters 3000.

it's really when you don't have these tools that you realize how good they are... (it comes at a price though)