Flux Capacitor - by Back to the Future | Trading Reviews and Vendors


futures.io - futures trading strategies, market news, trading charts and platforms


Trading Reviews and Vendors


Discuss and review vendors of commercial trading products, trading rooms and services, trading indicators or third-party paid add-ons




 

Flux Capacitor - by Back to the Future

  #28 (permalink)

fairfield, ct
 
Trading Experience: Advanced
Platform: Ninjatrader, InvestorRT
Broker/Data: Vision, Rithmic, IQfeed
Favorite Futures: ES, 6E, CL, NQ
 
Posts: 8 since Aug 2009
Thanks: 41 given, 29 received

possible ingredients to Flux Capacitor predictions....

Hi y'all,
I have a theory on the software ingredients that comprise the Back to the Future prediction engine.
1) Neuroshell Day Trader Professional Software
This software can be found at Neuroshell (can't post links yet, only first post And this link, shows several of their models that used neural network & genetic algorithms to predict price sometimes 5 days in the future based on the software studying past data:
Look for their link of GeeWhiz examples under "Product Info"
2) There are many available plug-ins to Neuroshell, which Neuroshell can use in it's strategy models and prediction models. The plug-in list can be found here:
under "Product Info", under Add-ons
3) There is one plug-in company whose indicators look remarkable like Back to the Future's AMACD and Power Zone, to my eyes,...the company is called Icinder. Their website is:
Icinder
4) I theorize that the AMACD is actually iCinDER™ V2, link to desciption at their site with chart with V2.

What can be done with Neuroshell and this indicator is Neuroshell can analyze 50-60 days of price data and predict the V2 indicator forward hours and/or days.
5) I theorize that the Power Zone is actually iCinDER™ W3, link to description here:
at their Icinder site
image here:
also at site
Notice, how the W3 indicator is tracking longer/larger cycles/moves and the V2 is tracking price more closely like the AMACD. Again, Neuroshell can analyze W3 over years or months of data and predict W3 hours, minutes, and/or days in the future.
Icinder's help manual with a little more information can be found here:
look up at the Icinder site the link "Icinder for Neuroshell", then you can find a link that says something like "you can access help files for v2 and w3 here"
Go through their tutorial and you'll see some prediction examples with Neuroshell and with Neuroshell's other software package for trading as well as science apps called "Chaos Hunter"

I believe that prediction algorithms could be discovered within Neuroshell with W3 and V2, perhaps different prediction algos for different markets, then coded up for Ninja Trader or other platforms, rename the indicators, but keep the code and prediction algorithms, then supply Ninja Trader with historical data and let the algo + indicator analyze and then predict their plots a certain amount into the future.

Obviously, there could be more inputs/variables that went into the prediction algorithm, but I'm pretty confident that I'm on the trail here. There are a few other well known AI prediction trading software analysis/forecast packages out there: Wave59 has one, Mark Jurik has his own version, Biocomp Systems have 2 packages...I believe their package Profit 8.0 has this predict ability-could be there other one-and there are other lesser known ones...such as Salford Systems MARS-Multivariate Adaptive Regression Splines at Salford Systems website...which is supposedly the super duper industrial strength engine for such predictions...which could also be fed similar Icinder cycle indicator formulas with price data to produce prediction formulas. They could have studied Icinder or like formulas with these packages and then written their own more portable neural net/genetic algo for pluging into Ninja Trader/Metatrader...that's my theory.

I think it's very interesting how AI can help us in analyzing the markets...Neuroshell is a pricey package, but, if a group of trading buddies all pitched in together, it could be a fun toolset to create strategies and prediction models that a group could share. Another Neuroshell plug-in company that intrigues me is Noxa Analytics at "NoxaPredict"

there's a Noxa downloadable tutorial and pdfs at their site, just snoop around for the Neuroshell Noxa packages, and you might have to give them name/email to get access to some of the pdfs...

here's a Noxa link to their Entropy indicators/tools: somewhere at their site

Again, the important thing here is that all these tools inside of a package like Neuroshell can study past price and predict themselves into the future. And, I believe you could then either code up or buy the rights to neural nets/genetic algos (i.e. Jurik's engines) and port them into Ninja Trader with the goal of predicting certain indicators that you could also buy the rights to & rename or code up something similar. Curious to hear your guys feedback on this theory.
-happy near year and best wishes always, I go sleep,
BG

p.s. from my studies in ES, I'd say the reason Flux doesn't work as well is because the ES has many more institutions battling each other's positions and has greater volume thus greater numbers of players interacting with eachother, making it harder to predict with such models, BUT easier to predict using simple volume tracking methods like VSA and cumulative delta bid/ask volume indexing, etc, just have to remember that the cycles are often jumpier & quicker for intraday, that's the trade off for all that liquidity...if someone studied the intricacies of volume cycles of the ES, I believe the prediction models could be better tuned to provide more accurate readings. But, it makes sense that less violent, slightly to greatly less liquid contracts that trend a bit longer/smoother would show up better under the prediction models...and obviously it makes sense that contracts that represented something real/natural like a commodity or even a currency may have more natural cycles that could be picked up easier than the battles going on in the ES.

Reply With Quote
The following 9 users say Thank You to bgtrader for this post: