2. Member Dashboard Guide
Once you have joined Crypto-ML as a free or paid member, you will spend most of your time on the Member Dashboard. This document provides an overview of how to use the Member Dashboard.
The Bitcoin and Ethereum price predictions are the cornerstone of Crypto-ML’s platform. These predictions are generated by our neural networks.
- Price is predicted both 6 and 12-hours from now.
- Predictions are updated every 15 minutes.
- The predictions change automatically on the Member Dashboard.
In this example, you can see:
- Bitcoin price is predicted to drop -4.7% 12 hours from now.
- Ethereum is predicted to trend up over the next 6 and 12 hours.
Note that the predictions are not meant to be interpreted alone. Please see the interpretation notes below.
► Learn more about our crypto price predictions
In addition to price prediction, Crypto-ML machine learning also identifies when the market is not following traditional patterns. This occurs rarely but can result in huge market movement.
Manipulation detection shows below the price predictions.
We know our machine learning models predict accurately a certain percentage of the time. When the predictions stop tracking with reality, that means the market is behaving differently than it normally does when all of the current conditions are true. When this happens, we no longer have a “normal market.” Instead, we have a “manipulated market.”
The indicators can be read as follows:
- Gray denotes a normal market. The predictions are expected to be correct.
- Green denotes bullish manipulation. The market is expected to go up more than predicted.
- Red denotes bearish manipulation. The market is expected to go down more than predicted.
► Learn more about our manipulation detectors
Interpreting Predictions and Manipulation
By having both the predictions and the manipulation detectors, you can now effectively navigate what our machine learning is telling you.
|Positive Prediction||Negative Prediction|
|Normal Market||Bullish: Follow Predictions||Bearish: Follow Predictions|
|Bullish Manipulation||Strongly Bullish||Bullish|
|Bearish Manipulation||Bearish||Strongly Bearish|
► Learn more about interpreting predictions
Crypto-ML will generate trade alerts based on market conditions, predictions, and manipulation detection.
When this occurs, you will see the Member Dashboard section update accordingly.
- Trader members will receive notifications when the trade status changes.
- Auto Trader members will receive notifications and have trades automatically execute on their behalf.
To understand the difference between the various trading pairs (BTCUSD, ETHUSD, USDBTC, and USDETH), please see Choosing Your Crypto Pair.
► Learn more about trade alerts
While Crypto-ML’s other indicators are focused on short-term timeframes and trading, the Market Index is designed to provide a big-picture view of markets.
The Market Index uses big data and machine learning techniques to measure market sentiment and determine a score.
- If the index exceeds +80, a bull cycle may be nearing its peak.
- If the index goes below -100, a bear cycle may be ending.
Note: Major market trend reversals have only have happened a few times in the past. Therefore, it is difficult to apply confidence levels to the Market Index output. It should be used in concert with other indicators to help identify trend changes.
► Learn more about the Market Index
The Signal Management section is designed to give some insights into our machine learning models.
- Goal to Maximize: supervised machine learning models have goals. Here you can see what goal the model is given. In this case, we are either seeking to maximize the quantity of fiat (USD) we hold or to maximize the quantity of crypto (BTC or ETH) we hold.
- Release: covers the version of the model currently in production. The release is also noted historically on our Trade History.
- Prediction Model: described the machine learning algorithm type used.
- R-Squared: at a high level, this describes the percentage of predictions that fall within a standard deviation of reality. In the picture below, the BTC/USD predictions are right 94% of the time within a +/- 0.33% range.
- Mean Absolute Error: the standard error of the predictions.
R-Squared is our key measure for model effectiveness.
► Learn more about the R-Squared vs Trading Effectiveness