Merge pull request ad-freiburg#6 from JoBuRo/master

Added post about my bsc project, BTC trading app

content/post/project-btc-trading-app.md

@@ -0,0 +1,168 @@
+---
+title: "Bitcoin Trading App"
+date: 2019-10-04T13:42:55+02:00
+author: "Johannes Herrmann"
+authorAvatar: "img/project_btc_trading_app/portrait.jpg"
+tags: ["bitcoin", "crypto", "trading", "trading algorithms"]
+categories: ["project"]
+image: "img/project_btc_trading_app/btc-background.jpg"
+draft: false
+---
+By anticipating market movements ahead of time it is possible to generate
+profits. This application implements several trading strategies which aim
+to do that automatically.
+<!--more-->
+
+# Content
+
+1. <a href="#introduction">Introduction</a>
+1. <a href="#chart">The Chart</a>
+1. <a href="#traders">The Traders</a>
+  1. <a href="#gen">General Trader Settings</a>
+  1. <a href="#bnh">Buy and Hold</a>
+  1. <a href="#ma">Moving Averages</a>
+  1. <a href="#macd">MACD</a>
+  1. <a href="#stochs">Stochastics</a>
+  1. <a href="#rsi">Relative Strength Index</a>
+  1. <a href="#stdev">Standard Deviation</a>
+
+# <a id="#introduction"></a> Introduction
+
+Since markets exist people have been trying to anticipate changes in the price
+of an asset ahead of time, in order to generate profits. In theory it is simple:
+buy when the price is low and sell when the price is high. But whether the price
+is actually high or low depends on future prices, not past ones. A price is
+considered high, if it is going to fall in the future, vice versa it is
+considered low, if it is going to rise.
+One of the tools
+for predicting market movements is Technical Analysis (TA).  
+This application implements some of the most promising techniques from TA
+for trading crypto currencies (e.g. Bitcoin). These techniques are then used
+by a "trader" to predict future price movement and generate Buy- or Sell-signals.  
+This application also provides options to adjust parameters of techniques
+and to combine them to create a trader that produces more accurate signals.
+
+This blog post will explain how to read the candlestick chart which displays
+the price and trader actions and how the implemented TA techniques work.
+
+
+# <a id="#chart"></a> The Chart
+
+<img src="/../../img/project_btc_trading_app/candle.jpg" width=80%>
+
+A common way to display an assets price over time is a candlestick chart. Each
+candle represents price action over a certain period of time, also called
+"granularity". Often a value of one week, day or hour is chosen. A candle
+consists of a thick body and two thin "wicks". The body shows the opening and
+closing price in the selected time period. If the candle is red, the price fell
+from the opening to the closing value and if the candle is green, the
+price rose from open to close. The upper part of the thin wick represents the
+maximum price over the period and the lower part represents the minimum.
+
+The solid lines in the chart show the values of the trader accounts and can be
+toggled on or off by (un-)checking the box "Show trader account value". The
+green line represents trader one's account and the blue line trader two's
+account.
+
+Under the tab 'General Settings' you can choose the asset, the granularity (size
+of the time interval covered by one candle) and whether to use historic or live
+prices.
+The historic prices are chosen by selecting a date range. The live prices are
+selected by entering the number of candles (units of time) that should be
+displayed.
+
+# <a id="#traders"></a> The Trading Strategies
+
+## <a id="#gen"></a> General Trader Settings
+
+All traders regardless of their used strategy have several general settings:
+
+### Stop Loss:
+
+This percentage defines how much value the asset can loose before the trader
+will sell it. This is a simple form of risk management.
+For example: If the trader buys at 1000 USD per unit and stop loss is set to
+10%, it will sell everything it bought when the price falls below 900 USD per
+unit.
+
+
+### Take Profit:
+
+This percentage defines how much the price can rise before the trader secures
+its profits. This is a sort of risk management, too, like stop loss.
+Example: If the trader buys at 1000 USD per unit and take profit is set to 10%,
+it will sell 100 USD worth of asset when the price hits 1100 USD per unit.  
+Disclaimer: This mechanism is based on the premise that the USD is more stable
+than the asset price, which is true for most crypto currencies.
+
+### Fees and Budget:
+
+Fees are specified in percent-tenths (0.1%). In practice, they can range from 0.5% to about 2%, depending on
+the used crypto exchange.
+The trader budget can be specified by a starting value, which the trader will
+recieve right away. And a regularly added amount can be set, along with
+a frequency defining how often the trader will recieve it.
+
+## <a id="#bnh"></a> Buy and Hold
+
+The "Buy and Hold" strategy is the most simple one. Whenever there are USD in
+the account, use them to buy as much of the asset as possible and hold it.
+With this strategy, bought assets are never sold.
+
+## <a id="#ma"></a> Moving Average Crossover
+
+This strategy is based on two moving averages. One takes a large sample size and
+is called "slow", the other takes a smaller sample size and is called "fast".
+In theory, the slow moving average is an estimation of the long-term price and
+the fast moving average estimates the short-term price.  
+
+A Buy-signal is triggered, when the fast moving average crosses over the slow
+moving one. Vice versa a Sell-signal is triggered, if they cross the other way
+around.
+
+In the app you can choose between two "flavours" of moving averages: simple or
+exponential. The exponential moving average puts more
+weight on recent data points while the simple one assigns the same weight to all
+of them.
+
+## <a id="#macd"></a> MACD
+
+This strategy is based on the MACD: moving average convergence-divergence. It is
+calculated by taking the difference between a fast and a slow exponential moving
+average (EMA, see <a href="#ma">Moving Averages</a>). Then, the average of this
+difference is computed, which gives the MACD-signal.  
+If this MACD-signal crosses above zero, a Buy-signal is triggered, if it crosses
+below, a Sell-signal.
+
+## <a id="#stochs"></a> Stochastic Oscillator
+
+This oscillator's value is computed by first looking up the highest and lowest prices
+in a defined time period. Then the difference between the current and lowest
+price, as well as the difference between highest and lowest are calculated. The
+ratio between those is then compared to a moving average of itself.  
+If the ratio
+crosses above the average, a Buy-signal is triggered and if it crosses below the
+average, a Sell-signal is triggered.
+
+## <a id="#rsi"></a> Relative Strength Index
+
+To calculate the relative strength, upwards and downwards price movements are summed up
+respectively and averaged over a given period.  
+The relative strength index is the ratio between the average of upwards movements and
+the sum of both averages.
+
+There are many ways to interpret the RSI and there does not seem to be one
+agreed upon interpretation.
+For this app, the crossover of the RSI with an average of itself is considered.
+If the RSI crosses above the average a Buy-signal is triggered, if it crosses
+below the average, a Sell-signal is triggered.
+
+## <a id="#stdev"></a> Standard Deviation
+
+This strategy simply calculates the standard deviation of the price and a moving
+average of the standard deviation. Contrary to the other strategies, this one
+does not produce Buy- or Sell-signals. Instead, it signals whether a big price move
+is imminent and thus if the trader should act or not.
+If the standard deviation crosses above its average, a relatively big move is
+expected and the trader can act. If it crosses below its average, no move is
+expected and the trader can not act.