Market Scans

[trackers]

Any TA'ers out there use market scans in their TA software?

I'm experimenting with some breakout / trend reversal stuff but still learning at this stage..

[wbosher]

I use Real Edge sofware from http://findata.com/

You can scan for Medium Term Uptrends or whatever you're after, and then eyeball the results to look for chart patterns. The subscription is about $20 a month for one exchange (NZX, ASX etc...) and the software is free. It's not the best on the market, but for a newbie it's quite cheap while you're learning.

[Phaedrus]

The NZ market is so small that no stock screener is necessary.

The Aus market is rather bigger and you really want access to some sort of scanner. The simple one bundled with IncredibleCharts is all that most people would need.

With over 10,000 stocks, US markets necessitate the use of sophisticated and powerful scanners. I use the ones at TD Ameritrade.

[trackers]

I use Real Edge sofware from http://findata.com/

You can scan for Medium Term Uptrends or whatever you're after, and then eyeball the results to look for chart patterns. The subscription is about $20 a month for one exchange (NZX, ASX etc...) and the software is free. It's not the best on the market, but for a newbie it's quite cheap while you're learning.


Hi Wbosher.. That's not a bad deal, I take it thats end of day ASX data?

Phaedrus you don't use Metastock 'The Explorer' scanners?

I find it hard to cut through all the noise - The scanners bring up Warrants, consolidating shares and things quite often. I think I might need to create a list of every share on the ASX that I'd be willing to part my money towards... But thats a big ask!

I also really need to get a decent broker + realtime data...Interactive Brokers is proving to be quite a mission to fund (plus I'm behind a firewall so can't access its advanced trading platform)

[wbosher]

Hi Wbosher.. That's not a bad deal, I take it thats end of day ASX data?

That's right, it's EOD data. I used it for the ASX, but I believe the price is the same regardless of what exchange you use. Obviously you can subscribe to more that one exchange for a little more $$$.

[Ponda]

Trackers:
I am experimenting with ProRealTime which is an EOD data which is free.
I find it educating and am using it for that.
It shows Boll Bands, resistance, supports etc.
It doesn't do the NZX but does do ASX and a truckload of other exchanges as well.

I don't have the self discipline to use it on a regular basis. (It was one of my News years goals, but feel that I may have failed on that one)

I have attached the link.

http://www.prorealtime.com/en/

[trackers]

Cheers Ponda that looks pretty awesome might look into that. Anyone here use bollinger bands as a primary entry signal? Back-testing 100 days worth of ASX data on 70 select stocks using common indicators shows bollinger bands come out on top

[peat]

trackers
can you explain what your criteria were for that bolli system? I've seen systems described with bollingers as fading the 2 sd upper channel
http://www.investopedia.com/articles/trading/07/bollinger.asp
but others seem to buy the upper channel , how did you work yours?
http://www.investopedia.com/articles/trading/05/022205.asp

[trackers]

trackers
can you explain what your criteria were for that bolli system? I've seen systems described with bollingers as fading the 2 sd upper channel
http://www.investopedia.com/articles/trading/07/bollinger.asp
but others seem to buy the upper channel , how did you work yours?
http://www.investopedia.com/articles/trading/05/022205.asp

Hey Peat,

"This is a good example of the cross() function and the top and bottom Bollinger band functions. This system enters long/short when the close has been below/above and then crosses above/below the lower/upper Bollinger band. Long/short positions are closed when the close crosses above/below the opposite band. "

Except I disabled sell short orders as there's no point doing a sim of a trade you can't actually perform in real life. Might extend the back test out another couple hundred days and see if the results are the same... Thanks for that link AA, guess that makes pretty good sense. If I can short-list a few more indicators like this one to use when things start getting choppy that'd be a good result!

[trackers]

I've been trialling a system which checks that the closing price of a security is above a moving average (moving average period optimised by backtesting). It also checks for 4 out of 5 crossovers on a point system to have occurred within the preceeding 3 days:

RSI crossing 50 from bottom (period optimised by backtesting)
Stoch crossing 50 from bottom (period optimised by backtesting)
RMI crossing 50 from bottom (period optimised by backtesting)
PDI crossing MDI

Its actually given quite an average performance, which is highly disappointing... Think I need some volume indicators as well, and maybe some longer term trend indicators.

A scan I've come up with which has returned a lowish profit but an amazing trade profit/loss ratio (consistently around 2:1 for most of 150 stocks) is entering a trade when closing price crosses through the bottom bollinger band, and exits the trade when the price drops below an optimised moving average...

I found that bollinger bands are good to get into trades, but sell signals are few and far between. The system is one of the few out of the 100-odd I've seen that shows profit more often than it shows loss; most are the opposite but the profits are usually larger than the losses which makes up for it.

[Phaedrus]

A lot depends on what you want from your scanner. Initially, my aim was to to create a scan such that when run, it would come up with a single stock which I then bought.

Later I found it best not to try and make my screens too good. In other words, if I made them so tight as to filter out all unsuitable stocks, they also filtered out some good prospects. I now aim to end up with a list of "possibles" (rather than "probables") and then flick through their charts. You can't beat the Mk1 eyeball! It doesn't take long to work through a "short" list of as many as, say, 50 stocks because each one takes only a few seconds.

100 days is a very short backtest period, Trackers. Results obtained would probably not be all that robust. Ideally, the backtest period should be long enough to cover uptrends, downtrends and trading ranges. More really is better in this instance!

[trackers]

Hey Phaedrus,

Thanks for that - Mk1 eyeball, I love it!!

I've been struggling for a while whether to backtest over the short term to identify systems that are working right now, or over the long term at systems that have prevailed over a number of years.. But of course you're right, with a long term view I should be looking at long-term systems.

The longest scans I've run to date have been 600 days on selected stocks and Stochastic 20/80 system has recorded the highest overall average profit, but at a profit:loss ratio of less than 1:2 (which I take to mean the system returns two losses for every one profit).

All fun and games, I really enjoy doing this which is a big bonus (just need to get better at it)