Topic: Haasonline Simple Trade Bot For BTCe and Bitstamp[Main topic] - page 113. (Read 416703 times)

Sean, are you using 0.9 alpha or 0.8 - I'm still using 0.8, I didn't get on board soon enough to get in on the alpha testing.  When you say minimum drop settings, are you referring to the profit and drop safeties?  I think I'm coming to the same conclusion, the closer these are to the exchange fee (ie 2 times instead of the 6 to 12 mentioned in the setup instructions), the more likely it is to execute an order that ends up profitable more often. 

While watching for a few hours, using RSI 30 with 30/70 and 60 second interval, a setting at 2 times the rate would have executed 3 times with a profit of approximately 0.4% for each trade, while a setting at 5 times would have executed a buy order, but never a sell.

Would love to compare notes and see how different approaches affect the results.

Thanks for the screen cap enki74, but what are you using for settings?  RSI signals?  At what threshold?

Don't answer if you are of the opinion that this type of information should be kept "secret".  Many feel they have the "secret formula, and are afraid if they share it too many people will learn it and it will not work anymore.

This position is not well thought out.  Trading has been around so long, there are no "secrets".  There are only informed decisions & strategies.  I learned everything I know from other people, and will in turn pass that to anyone who wants to learn.  We all learn from each other, and IMO it makes us all better traders; especially is a small community like this one (STB).

That said, I am currently testing the RSI signal generator.  I want it to buy/sell fast on sudden, pronounced movement (spike) but not necessarily on more gradual movement (trend).  I also want it to recover as soon as possible from an 'out of trade range' condition.  Toward that end, I am using RSI30 with 10% & 90% thresholds and the minimum drop settings allowed.  The bot is making decent trades with these settings, but as I learn to use it better, I hope to refine that.

Pretty sure he said that a portion was left behind with purpose ....

I'v a question/remark about version 0.8 (latest)

I wanted the bot to sell all my bitcoins i had with the high/low bot and selected the all in option.
After a while the boundaries where reached and the trade was made. I got however stuck with 0.0073 BTC, for which no sell order was made.

A closes look showed me that not the complete amount of bitcoin a sell order was given for, but a portion was not sold, for which i assume the fee had to be paid off.
But either the calculation is wrong, or the fee is paid from the sales value, and no reservation should have been made for that (when trading manula on the btc-e, If i would sell all, i can sell all and fees are just subtracted from the sale earnings)

Is this bot behaviour done by design, or did i find a bug?

So here is what mine look like... http://s1329.photobucket.com/user/enki74/media/trading_zps9f479024.jpg.html

Oh thank God lol.  I was truly at a loss how to explain it differently.

Please do not be offended by my manner, I confess I am more a "numbers person" then a "people person".

With that basic fundamental out of the way, lets get back to the propose of this forum; Stephan's bot.

Just so it is clear, I am just a user/tester of the bot... have no other affiliation with it.  disclaimer on me, I am an experienced trader, but pretty new to playing with bots; I am learning.

One of the things I like about Stephan's bot, is that it warns you when you are making a bone-headed move on your settings; then actually does a pretty good job protecting you from both the market and yourself.  I am a great tester of these features, as I am a master of bone-headed moves.

So... back to the question that started this whole thing, trade amounts.  We now know, from the math, that "bigger is better" when trading.  So:

Step #1.  Using smaller amounts, get a feel for the market and where your profit points are.

Step #2.  Use as much money as you can afford to risk in making your trades.

If you succeed with #1, and your trade strategies are valid, you will make the most money with step #2.  If not, you will lose the most lol.

That said, I would love for there to be more discussion in this forum geared specifically toward settings & strategies in using STB.  This is _not_ a simple "plug & play" bot, despite it's name, & I would love to get a better grasp on using the tools it offers.

Edit: -

$1000 traded will require fee of $4 correct?

$1 dollar traded will require fee of .004 correct?

If you could please explain that market does not need to move more to cover the cost of $4 than the one of $1 dollar

O shit, I got it from putting it this way,  what your saying is one makes more $ on the bigger trade to offset the cost.

Got it, perhaps you could have said that?

Yes let move on to some good stuff now I had to get past that :/

OMG inki74 do you not see the line that reads "Required market movement: 0.004 PPC" for BOTH amounts?!

For the love of God man!  

MINIMUM MARKET MOVEMENT DOES NOT CHANGE WITH AMOUNT!  

See in line #5 in Stephan's explanation?  See what is says?  See "Required market movement"?  What is the amount?  0.004 PPC, whether you are trading 5 PPC or 30 PPC.  Right?

Sir, if you truly cannot separate _market movement_ (does not change with trade amount) from _trade cost_ (percentage of trade amount, does move with trade amount) then trading may not be a good idea for you.  You are completely failing to grasp the most basic principals necessary to do this; even when put before you in the most plain English I can muster.

I wish you luck, but fear I have failed to help you.  I am not going to keep beating this poor dead horse.

So using your example when

Trading 5 PPC + Exchange cost of .02 PPC
Trading 30 PPC + Exchange cost of .12 PPC

In order to make money on 1st example market must move .02 PPC to break even
In order to make money on 2nd example market must move .12 PPC to break even

Do you agree? If not please explain.  In the example you give above you / by the number of PPC traded this implies that each trade is broken down into 30 individual 1 PPC trades in which case yes every trade would only need .004 PPC movement but this is not the case as we are even discussing what to set it to.

enki74,

________________clip________________
are you saying that it is wrong to think that one will have less trades while trading 1 btc at a time vs trading .1 btc if not then Im still not sure how this works because from my math while trading 1 btc X 2 trades @.2% each is .004 fee, and .1 btc X 2 trades @ .2% each is .0004 fee btc a smaller number than .004 and so my thinking is that the bot will be able to activate more often on smaller movements, if this is wrong then I'm really not getting it
________________clip________________

Yes, that is what I am saying... that is why I said it.  Yes, your math is wrong... that is what I said.  I agree, you are really not getting it.  Yes, that is sad, because I do not know how to say it more clearly then:

Your   trade   amount   does   not  affect   minimum   price   drop/raise.

Seriously.

Believe me or don't, but I gave you the formulas the whole trade world uses.

I can tell that you desperately want the rules to be different, but they are not.

You do make a good point of noticing that in a forum full of experienced traders, nobody is stepping up to tell you I am wrong.

So, given that Your   trade   amount   does   not  affect   minimum   price   drop/raise, if you like, we can move on to trade strategies (since this is how the discussion started).

We know that Your   trade   amount   does   not  affect   minimum   price   drop/raise.  Given this, we now know that there are only x number of profitable trades for any given time period.  Knowing that, we can deduce that your profit truly is tied to trade about; but opposite of what you were hoping.  The larger your trade amount, the larger your return, because trading locks you into strict percentages.  Lets do an exercise:

Let use our same round number examples and say that the market moves beyond our threshold limits 4 times today.  The price moves as follows:

8am  $100
11am $100.50
2pm   $100
6pm   $100.50

Providing your starting position was "in", you could buy/sell 3 times, because the market moved $0.50, and we established that $0.40 was our break-even point with a 0.2% fee.

It is important to state here that if these are the only times the market breaks your min raise/drop threshold today, these are the only times you are going to be able to make profitable trades today.  No amount o playing with creative math is going to change that.

So, if you traded using $1, you will make $0.001 twice... for a total profit of $0.002 today.

On the other hand if you have $10,000 in your account and trade it all, you would have made $20.

This, my friend, is why the rich get richer, and the poor fight at a disadvantage.  If your math were correct, it would be an equalizer.  Shame it is not.

@enki:
Lets take your example.

Trade amount: 5 PPC
Exchange fee: 0.2%
Exchange costs: 0.01 PPC per Trade    (5 PPC * 0.2%)
Exchange costs total: 0.02 PPC per Cycle     (0.01 PPC per Trade * 2)
Required market movement: 0.004 PPC      (0.02 PPC per Cycle / 5 PPC)

Trade amount: 30 PPC
Exchange fee: 0.2%
Exchange costs: 0.06 PPC per Trade    (30 PPC * 0.2%)
Exchange costs total: 0.12 PPC per Cycle     (0.06 PPC per Trade * 2)
Required market movement: 0.004 PPC      (0.12 PPC per Cycle / 30 PPC)

Exchange costs total = Minimum raise and drop warning level @ STBot v0.8.x

This is the simple explanation, the real one is more complicated because i need to take the exchange price in account for example. When i display a minimum value i need to calculate it back to the other coin currency.

I hope this helps.

I have already adapted the API of Bitstamp to support API Keys, this will be in the next release (before the end of the month) At this moment the old way of using the user id and password is still supported and working. (I have put a warning at my website about Bitstamp and there API)

The timer of version 0.9 is locked at 20seconds, i have confirmed that is bug in the Alpha release of version 0.9. However, just please only report the bugs of the Alpha version to me directly and not here. It has no use at all to post it here, people might get confused because of it. If the bug is in the beta release, you are of course free to post it here. But not for an Alpha release, most of the people will never see that version. An Alpha is made for testing, not for production.

------------------

Something else... i want to show the benchmark capabilities of version 0.9 Alpha. I already find out the Alpha testers are not really looking at this new feature, this is a shame because it is one of the best improvements inside version 0.9. Using the benchmark capabilities you can find out in detail what is going on and how to improve your trade bot and indicators on it.



Especially the Indicators tab is very interesting. It will show you all the signals the trade bot generates. The image below is a simple of a MACD indicator, but using the combobox you can also see the profit potential, extended currency information and all other signals.



Of course when we are done testing version 0.9 everybody will be updated and everybody can use these new features. (do not ask, just wait before the end of the month, it will come)

Hi,

Great bot, im testing with the latest version 0.9.0.0 alpha.

Unfortunately even though i have some problems and questions i shall leave them until i can use it properly. I can not use it until the timer setting is fix as its is stuck on 20 seconds at the moment and i cant really do much at that frequency.

I did also try to use it on bitstamp with just my login details but it would not connect and did in fact cause bitstamp to block my IP for 10 minutes (not quite sure how as you have to be polling really fast (1/sec i think) to get a block on your IP but i managed it somehow).

I shall give bitstamp another try and ask my some of my questions when the timer is fixed and i can test properly.

Excellent work so far , looking forward to tinkering with all the signals

Many thx.

Better enki74... thank you for separating your txt from the quoted txt... makes your posts more readable.

I apologize if I have not answered your question, I did try.

Let me try in a different way.  You are confusing the fee%/actualcost/movement%/movementactualcost ratios.

The basic flaw in your math is that you are somehow tying the trade amount to the movement/% fee equation.  That is simply not correct.

Using round numbers again, suppose BTC is worth $100/ea, & the exchange charges a 0.2% fee for each trade.

Because we know it takes two trades to make a profit, we know we have to double the fee (two trades) to arrive at what our minimum drop/raise will be.  In this case, it is 0.4%

This means we can calculate the minimum drop/raise thusly: 100*0.004 = $0.40

You have now calculated the minimum amount the price must move for you to make a trade and _not_ lose money.

Lets make a trade.  Lets Trade $1 worth of BTC.  If we buy @ $99.60 & sell @ $100 we have traded using the calculated $0.40 minimum, and have lost no money... end result = 0

Lets make another trade.  Lets Trade $100 worth of BTC.  If we buy @ $99.60 & sell @ $100 we have traded using the calculated $0.40 minimum, and have lost no money... end result = 0

Lets make a final trade.  Lets Trade $0.000001 worth of BTC.  If we buy @ $99.60 & sell @ $100 we have traded using the calculated $0.40 minimum, and have lost no money... end result = 0

Why?  Because (100-99.60)-(100*0.004)=0 (tradeamount[sell]-(tradeamount[buy])-(tradeamount*fee)
and (1-0.60)-(1*0.004)=0
and of course (0.000001-0.0006)-(0.000001*0.004)=0

This means that there is no trade amount that will affect the actual drop amount needed to make a profit.

Honestly sir, I am at a loss for how I could be more "forthcoming".  I grant that maybe I am not explaining it right, so I invite you to test in on your own.  Use small amounts tho, because math is a cruel, unforgiving master.  If you get it wrong, she will exact her price on you.



Sean

Ok whatever you don't seem to be very forthcoming with corrections....Im nothing if not good at asking questions none of which seem to be worthy of your time... Very superior attitude "As someone" who knows a few things about a few things and always willing to explain when I have a willing ear...

Hmm, my opinion is that a bot should be completely self sufficient.
I would like to suggest you add a clean-up procedure that checks your open orders that are no longer relevant and cancels them.

Otherwise if you leave the bot unattended for a longer period, you might have an unpleasant surprise.