Potential Neo Newbie

[lloydphilip]

It's also horrible at scanning and adding and deleting symbols. Apparently the developers think that it is a non issue.

[lloydphilip]

Annu,

What's the max number of symbols that you scan through in a day. Also what the max number of stocks you've automated in a day thru Neo?

thanks,
LP

[Annu]

Im not aware of a symbol limit (within Neo), but i only trade futures and FX spot, so i follow only a very limited number of symbols. I also dont do scans or simmilar. I think its best to ask these questions on the Tickquest forum, there are some very knowledgeable users over there, who usually chime in.

http://forums.neoticker.com/

[Annu]

P.S. there is a special 'Platform Conversion' section on this forum, and some threads there deal with migrating from TS to Neo. Worth a look i guess.

[peter_77_peter]

Dear Lloydphilip (or anybody else who could be of help),


>I'm a current TS user looking to leave as many of my equity strategies
>are based on multi-time frames. It's very hard to trade multis on TS as
>there are far too many limitations and ass backward things you need to
>do.

actually I´m facing difficulties with TS in the same area at the moment. My Trendfollowing Strategies take into account global trends derived from daily data and local trends derived from 60min data.

When I put this on different Charts, and communicate the Trend Signals via global .dll then I cannot backtest because TS will only actualize one chart and no communication with another.

When I put this on multi-data charts and play with data1, data2 and so on, I can backtest, because all data is available at one window, but: It seems that the order in that tradestation actualizes the data streams, the strategies, the indicators, the global.dll are not quite stable. That is, If I reload both symbols at the same time, it produces different results, than if i reload them one after another. Actually one of both results is perfectly matching my "different windows approach" so I know which one is the right choice, and probably it has to do something with the internal organisation of the interpreter with respect to multi data streams. But actually I don´t want to have to look for manual housekeeping of data consistency.

The other workaround that might work should be virtual candles, but actually thats a mess as well.

Besides that I´d love to have efficient money allocation between several symbols and strategies, and I don´t see a way to do this that is comfortable and backtest savvy....it´s basically the same lack of proper interfaces within TS and dirty workarounds with .dlls...

Actually sometimes I think that TS is a platform from the stone age, nice for prototyping but ugly if you really want to go professional.

I just wanted to ask you whether you´ve faced comparable difficulties and came up with a solution that might be of help in my case. Actually I´d love to leave tradestation, but first priority would be to get it fixed somehow without migration of the whole system into another platform.

best regards from spain,
peter

[lloydphilip]

Peter,

I'll be honest here. I programmed in Java about 7 years agao and hadn't touched any languages until about 1 year ago. When I started to work on TS, I thought my classes in C back in college (about 10 years) were more advanced than what I could do in TS.

Since I prmarily deal with equities, I have lots of new symbols everyday. I just can't create new charts or rows in radar to accomodate every 50 - 100 symbols everday. The data 1 and data 2 route seemed like a real pain. The fact is that my strategy requires analysis on the dailies, 30 min, 15 min and 3 min timeframes, and then looks for specific patterns on all those timeframes, it's just too cumbersome. In addition I would require 2 data streams to run my strategies. Hardly manageable in TS. So that's 4 for radar and 2 for charts.

Now I haven't touched DLLs let alone know what they stand for . After reading up on them along with the issues related to Data1, Data2 and backtesting with them, I said to myself that it's just not worth adding band aids and duct tape to what is broken. Not to mention my radar gives me phantom alerts right now which are quickly rectified by a Control R. I can can't imagine the number of phantom alerts that will come about from doing the whole data 1 and data 2 thing. Also god bless you if you run a strategy and refresh data. Kaput.

After doing a lot of research, I have finally settled on Neo. It does everything you want. But it does not look as pretty as TS nor does it have an easy interface as TS. It's kinda like going from Windows to Unix. Now if you leave the asthetics behind, you will have one last issue; data. For what I do, I like the TS data a lot. I don't need tick data. I just need 3 min charts at most. Most of my orders reside on the TS server and are usually sent way before the price tiggers. (Stop markets). Those bumbs don't have a proper way of doing stop limits. Go figure. Also, data, charting and execution are all part of one package.

Now that's where the simplicty and benefits of TS end. Neo on the other hand is 100 times more powerful. You can code in many languages, be it procedural or OO. In Neo it seems you can minimize your analysis down to minimum. Do you analysis on 1000 stocks on dailies, analyze the results which might be 200 on the 30 min, further analyze the resulting 50 on the 5 min and so on an so forth. If you program the darn thing properly it will filter out the unnecessary stocks and only anaylze what is pertinent. TS on the other hand analyzes and re-analyzes and re-analyzes. You need to program circles around it to stop doing each one of those things.

Also, in TS you need seperate functions for everything. Paint bars, show me, strategy, radar etc. In Neo you need one clean piece of code that will work for everything. In neo you can run you strategy from anywhere. In TS you have to run it from a chart window. In neo you can automate 20 stocks through one chart window, in TS you can't. Also Neo has a cool feature where you can view multiple charts in one chart. Really frees up the need for extra monitors.

Backtesting in TS is a true pain. If you want multi-timeframe and portfolio testing. You will need to buy inadequate products as add ons. In Neo you can do it all within Neo. Probably one of the best back testing pieces of software out there. The biggest problem is getting proper historical data. Now Michal has written some data exporter which can be used. But it seems to work one symbol at a time. Esignal only provide 3 months of historical, the unreliable and spikey DTN IQ provides 1 years worth. There are other far far more expensive options that will get you about 3 years of tick data.

Lastly, once you put all this together and climb the steeeeeeeeep learning curve, you will understand that your $1500 plus, $91 a month for esignal will be well worth it. Now based on some really crappy calculations, you would need to trade about 20 - 25K (roundtrip) shares a month to break even on the cost of Neo and eSignal vs TS. Hopefully you are taking only the best setups . Also, IBs exectuions are much better that TSs. Since I can't do stop limits, I get slipped up at least 5 cents rountrip on each trade. Keep in mind I trade some of the more heavily traded stocks out there. So in that sense I'm paying at least 7 cents a share with TS. I might only pay 3 - 5 cents in IB. The problem is that TS is like crack. You need more and more once you get your first hit. You know it's not good for you but you stick with it out of convenience.

That being said, TS still is a much better solution when compared to most of the other brokers out there. Hope that helps.

[peter_77_peter]

Hi llyod,

thanks for your prompt and honest answer. Actually if you mess around with TS in Detail you see that this is an architecture Problem of Tradestation aiming to be an easy to use interpreter, leaving professional users unsatisfied with the lack of global data management, proper interfaces, objects, events, and everything else you basically need if you want to diversify strategies, time frames, portfolio......I think Neo should be good at all this, because the approach is right, giving you interfaces for data and order management and leaving you lots of freedom (and thus responsibilities) to do whatever you have to in between........

thanks for your detailed explanations,
best regards,
Peter