OmniTrader questions

[skiabox]

I've been using omnitrader some time ago and I have a couple of questions.Suppose I se the trading system to Stocks (short term aggressive).Then some results are shown in the focus window.If I click on a stock and then push chart the chart windows appears but the signal of the stock frequently changes!If I run the system again the stock remains with the newer signal!What exactly is going on?Is there a site with omni hints & tips?

 

[Jazz]

In doing an Omni pick every day, I have found that some new longs have miraculously changed to new shorts a few days prev & vice versa. I dont know how these discrepancies can be got rid of. A few people have contacted the company about the lack of success re the signals but as far as I know, they havent received satisfactory replies.
If anyone reading this has had any luck with any explainations, please correct me.

Who did you buy your system from, you could try asking them for some settings?

Pop into the UK chatroom sometime during the week, there are people there (me included) who can show you their settings.
I use short term aggressive, with a few non-optimised settings changed, so they don't make a diff to the voted signals, but when you display the chart the indicator shows the diff setting (I think).
I set the filters to only show ADV & BTHR of over 80 for the point of the daily omni pick.

I dont use my Omni to chose stocks for me, but I like the amount of different things you can have on the chart at the same time ie fib, trendlines, supp & res, volume, & an indicator. As a timesaver its great, as a stockpicker, hmmmmmm.....

Try an Omni search to check out others opinions on Omnitrader on this BB.

Regards
J

 

[darth trader]

I bought mine from http://www.signaltraders.com and Chris Byars is the UK representative. There is a private bulletin board on the site that contains detailed instructions on how to fix the problem...

as I recall you need to create/copy some daily moving average cross over systems for 50 100 and 200 days (or some similar variant).

I agree with your view that it has excellent charting capabilties but I'm not a fan of consensus signals for picking stocks. My best success with consensus signals is with just a small number trading systems.

I prefer to use the metastock explorer to identify buy/sell candidates.

PS. the actual value pairs are:
8 and 15
21 and 55
70 and 200

all non-optimised daily.

Good luck,
Darth.

 

[darth trader]

More on omnitrader signals changing.

Of course the signals will change whenever the backtest is run to pick the best systems. I assume that this is not the problem that you are experiencing...

Darth.

 

[Jazz]

on signalwatch site, click on Omnitrader & FAQ

It recommends that you set the program to back test every 7 to 10 days, maybe longer if you are holding multiple positions. If you have backtesting set for 1 day, there will be large changes in signals, simply because of current market volatility.

Hope that helps
J

 

[titus-uk]

I agree, in this current climate, signals on OT can change wildly from 1 day to the next - mainly, of course, for TMT's. As a charting tool, it is excellent.

I am still experimenting with the settings. One way to try and bring stability is to edit the model by delaying entry until the price rises above a MA (eg 15 days). This seems to eliminate some of the volatility. Seems to work so far, but wil let you know how I progress/ experiment further.

Cheers,
Mark

 

[darth trader]

yet another thought..

you can set the program so that stocks with open positions are not included in a backtest.

 

[skiabox]

my post no2 !

What amazes me most is that after the signal has been changed from the chart display way it can't change back to the initial setting even if I run the test again!The change is permanent for that day!(I use end of day data)

 

[hi5]

I spent 3 months trying to get to the bottom of why signals change, what the optimum backtest setting should be and the effect of changing trading model on signals.

I discussed the backtest setting at length with US tech support. They advised 20 days. I put the following argument to them:

"Unless there is a dramatic change in the nature of a particular stock since the previous day, the systems used and the vote line should remain the same if backtesting is done daily. Increasing the time between backtests increases the probability of change in the nature of a stock and a consequent change in systems used to generate the vote line. I would expect
backtesting every day to "smooth out" the dramatic changes seen when backtesting is only done periodically."

They agreed - 1 day is best!

The reason why signals change when you select a different trading model is because of the effect on entry and exit conditions. Changing trading model will affect the capital return produced by each of the 120 trading algorithms over the backtest period (default 250 days). The 6 best algorithms are then selected to create the vote line.

Despite my efforts to optimise the trading model, I became very frustrated by the candidates offered to me by OT. A quick eyeballing with my very limited TA knowledge would eliminate most of them, most of the time.

I became very concerned upon hearing that Ed Downs had suggested that OT might not even work for the UK and Australian markets.

After 3 months of trying to get OT to give me reasonable trading candidates I gave up and returned the product.