Topic: (Closed) Butter Bot!: Premier Bitstamp, and BTC-E EMA Trading Platform (Closed) - page 57. (Read 274813 times)

Pablo, how is the bot downloading the trade data? I've been working on something and would love the raw data it uses to create the EMA output.

They're not /that/ interesting once you factor in the hefty btc price jump variance. Remove that and you're left with 30-40% trade profits, which many other settings are also providing. They're good, I'm just saying it's not a goldmine unless you did this exact thing starting over a year ago.

If btc holds at current values for the next year (hypothetically), for example, the trade profit would be single digit.

It's a great example of the power of the bot long term, which is the point of it to begin with.

These are quite some interesting settings!

I think we should be looking at these more closely, the great part is that they are long term results on a long term liquid exchange, it's really quite interesting and quite an improvement on Buy and Hold.

If my numbers are right, you could have purchased 8.66 BTC with 100 USD at start which at last EMA trade would be worth 2,666 USD, but you are holding (also at last EMA trade) over 7,000 USD worth of BTC. Where is all this loss of BTC Butter causes?

Butter is a tool and it must be viewed and used as such. You need to invest time in developing your strategy and learn about the trading engine (EMA); our job is to provide you (the trader) with as many tools as possible to help you succeed and develop your strategy (altcoins, alternative strategy modules, back testing , etc.) but the onus is on the trader to put in the hours; just like in all trading .

It might be interesting if we could work on this automated settings adjustment idea and see what we come up with.



Pablo.

Hi Zubillica ,
   This is the reason we encourage you to back test 6 months or more, results over the short term, 1 week or one month are way too variable precisely because of these mean deviators we keep talking about, you will see more constant and reliable results over 6 months or more.

Please let me know if you need anything at all.

Pablo.

Hey lets break down what this bots settings actually mean.



The way I read the settings, it says to me:

This is so odd, I ran your numbers again and got a completely different result. Is it part of EMA that the current price is such a strong influence on backtesting? Or is there some other factor influencing these results?

http://f.cl.ly/items/2P111y293s2N3f0t1N1d/Screen%20Shot%202013-11-14%20at%2018.21.19.png

Also +1 on reporting in BTC

I've been having a go at trying out some numbers in backtesting.... Could anyone tell me if they're any good/advice?

http://i.imgur.com/Af10aLll.png

What COULD be implemented though, is a function where the backtesting calculates the average of all possible offsets.
That might be a more objective way to compare strategies.

Please somebody correct me if I'm wrong!

Yup that's the loving I'm talking about.

7900% beats B&H by almost 200%

Look at this thread: https://bitcointalksearch.org/topic/m.736808

I posted several sets of settings that beat buy and hold by a large margin.

All you guys have to do is setup EMA calculation in a spreadsheet like Excel, and run the exchange data into it. You could set it up so it shows many different strategies at once, and pick the ones which best suit your tastes.

Each exchange really needs its own settings because of the variation, so you could then choose one based on each exchange, and setup 1 bot instance per exchange.

I'm more interested in multiplying my bitcoins than my dollars so maybe this bot is not for me. I'm going to try the trial to see how it goes though...

Agreed, I would love to backtest several schemes against a down market trend and not just against a bubble like we have been for the last 2 months.

Since you say automated backend testing is resourceful (understandable ) - isn't it possible to download the "database" of the market over the last 1/3/6/12(+?) months and do the testing locally? That way all calculations would be done on the machine itself so it's not using all the resources of whatever server it is using for backend testing.

Just a random idea that might be handy..

It can be seen that way, and it can be seen that it makes you money. Both situations occur at times, you have to ride out the storm. I was skeptical for this reason at first too but once I gave it a shot for awhile and saw the patterns emerging, I was convinced and bought a license for the bot.

Since then it's paid for itself and tripled my small investment. I also hold bitcoins, as that's a good long-term strategy as well. Do both, there's nothing wrong with it. You can even trade by hand, many people do that. I can't because I'm not able to get into my exchange easily during work.

The settings are crucial, you have to find what you're comfortable with. I still try different ones out, and sometimes from it I lose, but the point is you can set the bot up in a conservative manner and you'll make steady profit, or you could be very aggressive and see what happens. Odds are if you're too aggressive, the bot will lose your coins.

I tried some more aggressive settings yesterday and overall I lost .07btc, and am currently re-evaluating my math. That's just part of it.

But, it's not such a terrible loss considering the .07btc which 'vanished', I never had to begin with a week ago anyhow, and I'm still "up" overall.

Some folks may not post their settings because if swarms of people started using them, it might make their settings less valuable.

Hmm this is worrisome if true. Less bitcoins is not a good thing.

And the fact is, if you just buy bitcoins for fiat, let them sit, sell bitcoins for fiat, you would have a much higher profit in fiat with that absolutely simple strategy than by using this bot (and pay in fiat for using it every month).

Hi Guys,
I'm observing Luicon and jbssm arguments and it's hard to disagree. Some of us play with BTC to earn more fiat but ther is a group (believe it or not of people who believe that fiat is long time gone and it's downhill from here (you saw this http://online.wsj.com/news/articles/SB10001424052702303763804579183680751473884... however useles it is post factum) so possessing bitcoins may have way more sense in the future.
Btw, should we believe Coinsetter saying that institutional and venture capital to stabilise growth http://blog.coinsetter.com/2013/11/10/bitcoin-price-explosion-isnt-bubble/ and it's not a bubble?


Concerning TSL, would it be possible to have a price displayed at witch your set TSL value will sell and switch of the bot?

Bendix