Topic: Grid Trading Bot for Ltc, Btc and other coins (Read 2464 times)

Great news guys !!!

Due to the latest volatility we had good profits on the Grid EA
Now the profits are not as big as many of you expect but for those who know these numbers are very profitable in long run

so let's see the update

LTC/USD  - "conservative mode" for 16 days 5 trades GridSize = $0.9

NMC/USD - "medium mode" for the last 10 days - 14 profitable orders, that is good so far for this pair, GridSize = $0.1

PPC/USD - "aggressive mode" for the last 7 days - 40 profitable orders, GridSize = $0.07 (changed it from $0.03 due to the high volatility)

Overall Profit of Portfolio so far: 5.6%

Conclusions:
When the price will move lower and lower we need to increase grid size so we have enough funds to maintain the DrawDown

Just an update on my bot

I've written it in php and now I placed it on 3 "pairs" and 3 different types of modes: conservative, medium, aggressive
Info: profitable orders - orders which were open on one level and close at 3 levels away in profit

LTC/USD  - "conservative mode" for 10 days no trade - it could be because there is not much volatility and the grid size is too big for this pair, but we stick to rules and let it run  GridSize = $0.9

NMC/USD - "medium mode" for the last 3 days - 4 profitable orders, that is good so far for this pair, GridSize = $0.1

PPC/USD - "aggressive mode" for the last 24h - 2 profitable orders, GridSize = $0.03, with this grid-size on profit I usually get 1.5% - 0.4%(fee) = 1.1% pure profit (3*$0.03/$6), and for now I think this is the lowest acceptable level, we need also to keep in mind that in case of a big down move here we could consume all of our funds to buy ppc


Conclusions:
In future will be much easier to use only 2 variables for this bot:
ProfitPercent - Profit generated from 1 profitable trade for only 1 Size of Order, starting from 0.5% for aggressive mode (note: keep in mind that to this value needs to be added every-time 0.4% transaction fee), this will also help to have less orders if price jumps quickly up, if price jumps quickly down we could use another value for ProfitPercent, but this is another story
Amount Per Level - basically the order size in $

K started beta for the bot

Called it AutoCoinBot , although now it is still not auto 100%, for the testing purposes I made a php script running on local server. I decided to go with a fixed amount of $ for each level - it is easier to track the profit and closing partial of order for profit when price rises.

here is an image of what I have so far

https://i.imgur.com/ywFAQnI.png

Red levels are the prices and amount I bought there (which means the order is still open - in reality I have that amount in my LTC balance)

Yellow levels are the so called pending orders, if the price will go there it will buy that amount for me

Blue level is where the price is now so I can check visual where it is

Now I am waiting for the price to either go up or down for at least $5 in order to see how all this works

Conclusions> There is a minimum amount for one trade(e.a. 0.1 LTC) and in future this has to be considered when using fee reduction algorithm(more details later). Since this is a long term strategy I do not expect to make a lot of money in the next week - but at least it will provide me the feel if I am in the right direction as backtests show. This strategy is somehow "bullet"-proof because in the end you will end everytime with either $ (more then you had) or coins. It is not like your balance will drop suddenly and you will not have anything

Yes, you are right, in real trading there is no need to sell and buy at the same time, that was for example purpose. The main idea in here is that you need to take your profits at the levels which you specify in the strategy(stick to your rules), imagine them as parallel orders, for example you made that sell in order to get your profit, now if the price comes back from that point of 860 to 800 again you buy one more time since on that level of 800 you do not have any open order. The main profit in here is done when prices fluctuates up and down which exactly happens these days.

I am still testing the best grid size and position size on each level, so far the best are at arround $20 grid size with profit on the 3rd level. Another approach is not to buy 0.1 btc at each level, but instead buy worth of $50 of bitcoins, this way on the 3rd level for profit you can close a small portion of your btc just for the profit, avoiding payint too many fees.

I'm unsure that I "get" this, why ever sell if you're going to buy at the same time?

If this relies on the price always going up, shouldn't you just buy 0.4 at 800 and hold forever?

Recently tested a grid - like strategy on BTC-e MT4 platform, easier for me since I know how to program on MT4

Strategy:
Use  BtcUsd for sample
Have an imaginary Grid of $20 distance between levels
Buy on every level of $20 and sell on the 3+ levels from it (this one was the most profitable from backtests - taking in account transaction fee)

e.a.
Buy 0.1 btc at 800
buy 0.1 btc at 820
buy 0.1 btc at 840
buy 0.1 btc at 860 / sell 0.1 btc from 800 (profit $6) and so on from every level during long term run

Necessary amount of money depends on the grid size and number of bicoins you want to buy at each level

if price comes back buy at each new level and so on

analyzed in theory and optimized the potential presets and found really nice results and already made some money trading it.
here is some of the backtest results (50% revenue in 2 months - even if I have 15% per month I will be more then happy)

https://i.imgur.com/iyz1k5A.gif

Risk: Not really much risk since at any time you will have either bitcoin or $ in your account, and if you hold long enough you will be in profit. And as we all expect the price of bitcoin to rise and rise we are on good track

Want to make a bot to start on btc-e to trade on any currency

Anybody tried similar strategies?