Topic: Chrome Browser extension: MtGox trading bot - page 23. (Read 72509 times)

10 and 21 days  if you use EMA 10 EMA 21 

Sorry if this sounds n00bish but how far back doest the period of the EMA take for its value in the bot 

I just want to preface this post by saying, I feel like a world-class idiot spamming up this thread with my n00bness...
Anyway... the above method doesn't work   Maybe it's fixed in the newest version of Chrome so that right clicking and choosing "save link as..." doesn't give the .crx file anymore???  Anyway, I'll just follow the link above and get the source that way but... If anyone wants to take the time to really dumb this process of retrieving Chrome Extension source code down for me, I'd pay 0.5 BTC...

OH cool, thanks... 
I had just found this, too:
"The first step is downloading the extension’s CRX file. Instead of clicking it’s link, which will lead to that extension being installed on your browser, right-click the link and choose “Save link as…”."
from http://orenyomtov.com/how-to-view-a-google-chrome-extensions-full-source-code.html

LOL, a durrr!!!  ::: facepalm :::

https://bitcointalksearch.org/topic/m.786714

I just don't see where you obtain the .zip archive -- I am likely just overlooking something very simple...

That's a BIG read but I'm looking into it and playing with changing exponential moving averages on charts too.  I don't know if this sounds stupid but I get the feeling the bot would do better if it updated the ticks to every 15 minutes instead of every hour 

and if i understand correctly there is a two hour wait after crossover even if you do set the threshold to zero, be nice if that were adjustable as well

Now it's getting a little bit confusing for me.

First I've successfully changed the MaxHoursBack value in Local Storage to 192. The results looked VERY different than before. The change from green to red has happened now 8 hours later.
Then I saw your additional EDIT regarding the MaxHoursToKeep value, which is 120 by default. I thought this could be the reason for the strange results, because I used MaxHoursToKeep<In the next step I set both to 120, but with the same strange results. After this I modified the refreshEMA function, so it always first sets emaLong = [] and emaShort = []. The strange results are still there.

I'm using the graphical output of bitcoincharts.com with the same EMA settings to verify the correct point in time for a trade. It differs quite a lot at the moment...

EDIT: Finally, I changed MaxHoursToKeep and MaxHoursBack to 48, just to learn how this looks, and the results are even more different than the expected output. Any idea?

Sure, changing this value is no problem. Thanks for your hints regarding the LocalStorage.

Indeed I'm using a threshold value nearly zero. I think I understand the threshold function, and I like the option to play with it.
However, wasn't it the recommended strategy in Goomboo's original thread to directly trade when the cross has happened, and wouldn't a (higher) threshold value delay the trading execution?
I prefer the faster trading reaction if the cross has happend, even if this results in more false-trading signals. But again, I really like the option and will look how it works out.
 

I don't think its quite possible. Unless you set the treshold to zero and had a very, very, very bad luck...

You have the source code - feel free to change the "48" in line 11 of background.html to whatever you want.
You can also go to the extension's "LocalStorage" (Inspect active views: background.html -> Resources / Local Storage) and manually add "MaxHoursBack" value there - that should work as well.
I was planing to make it configurable via the options, but then though that this was supposed to be a simple extension...

EDIT:
But if you want to always have exactly the same numbers, increasing the startup time would not really help here.
In such case you should rather set the MaxHoursToKeep (line 4 of background.html) to 48, instead. And would also need to change the refreshEMA function to always force the reset.

Thanks for your prompt reply. That sounds logically to me.
I could found it in your code.
I would suggest to change it to a bigger value then 48, if you even plan to do a update in the future.
The startup time was never a problem for me. Personally I could wait over 10 times longer until the tool showing data and start trading, especially if this helps to make the numbers more correct.

Another question came into my mind, which is maybe a little bit related to this issue.
I thought about possible worst case scenarios  which could result in loosing your trading value quite fast, because of unmeant tool behavior. I still worry about what could happen by running two chrome instances (on different computers). If this two instances have the same option setup, but are working with different numbers, based on the startup time, I worry about situations where the two instances are trading "against each other".  E.g. Instance1 would already sell BTC, Instance2 is still below the threshold, Marked price changing and Instance 1 is buying again shortly after this Instance 2 touch touch the threshold and is selling, or so. Especially if you run with a very low threshold setup. Do I just fantasize or wouldn't this also be a argument to make the calculation result, which is the trigger criteria, more independent from the tool startup time? Would love to read your opinion.

 

It's probably because of the number of candles which the extension has in the memory - and then calculates the EMAs upon.
At startup it fetches 48 candles (hours) - so the Exponential Average is calculated starting from 48 hours back.
Later, when you let it running and new hours are fetched - the EMA is calculated starting from more than 48 hours back...
But if you reload the extension, it again re-fetches only the last 48 hours.
Thus the difference - though, it shouldn't be really significant. At least as long as you don's use EMA parameter above 30 or so.

I could fetch more hours at startup to make the EMA values more precise, but then it would start longer...
48 seemed like a good compromise - but it's very easy to change it, if you think that longer startups would be better

As I said, it is a zip archive - unpack it with any un-zipper. Not with Chrome.

I'm wondering about why the historical EMAfast and EMAslow values are changing after a reload of the tool.

e.g.
Hour/Price/EMAfast/EMAslow/...
9:00/4.800/4.892/4.922/...
8:00/4.828/4.901/4.927/...
7:00/4.847/4.909/4.931/...
6:00/4.848/4.915/4.934/...
...

After disabling/re-enabling the tool it shows the table again. The historical price is for every hour the same like before, but EMAfast and EMAslow is quite different than before and therefore also the difference in %.
The difference in % is the value on which trades will be triggered. I guess it has something to do with the market price change, but I don't understand why this influence the historical EMA values?

Has someone an idea for this behavior of the tool?

Piotr, long time lurker here, I read all of goomboo's post before looking into this extension. Just wanted to say thank you and remind everyone else to LURK MOAR.

Sorry, just frustrating to see the difference in educated, thoughtful discussion and THIS SHIT IS LOSING ME MONEY.

Party on and thanks again for the work.

I plan to do just that if I can just get at the source code...

Wait -- spoke too soon.  I actually do need some clarity on this if you would please:  Please explain how you extract the source files?  I only see "pack extension" and "load unpacked extension"...  

wouldnt it be something to develop this kind of app for bitcoinica? :-D

nevermind, I think I've got it now, found the "developer" checkbox