Cindicator accuracy: high sustainable level reached
2019 accuracy summary: Hybrid Intelligence works
Today we are sharing the latest data from Cindicator’s analytical products. We’ll review our accuracy for 2019, provide examples and interpretations for indicators, and outline what’s next.
Progress since the token sale
2020 will mark three years since the first group of token holders supported Cindicator’s highly ambitious vision of a Hybrid Intelligence ecosystem.
Together, we set out to unite a community of decentralised analysts around collectively making forecasts and to turn this data flow into predictive analytics made accessible through the holding of tokens.
Our data funnel and product distribution goals have been achieved:
- The Collective Intelligence platform attracts and retains analysts who have made over 15,000,000 forecasts;
- 5,000+ indicators were distributed to 6,500+ token holders who use them in various trading strategies.
The overall accuracy is now consistently above 60%, validating the completion of both goals and establishing a solid foundation for further development. This robust figure is a measurement based on thousands of indicators sent to token holders over the past four years. Furthermore, this figure is free from any look-ahead bias because it only includes indicators that were published before the results were known.
2019 accuracy summary
Let’s review the data in more detail before you pore over it on your own.
Total accuracy has held above 60% for 5 quarters. Overall accuracy since the launch of CND-based analytical products in Q4 2017 stands at 61.77%. This, however, is the result of experiments which began way earlier in 2015, when the company was founded.
The accuracy of crypto indicators stood at 60.81%. This suggests that crypto indicators now perform well in any market environment, including bullish, bearish, and side trends for BTC and altcoins.
In 2019, the accuracy of fiat indicators was 70.52%. This broad category includes indicators for US equities, S&P 500, commodities, and various market events.
The number of EPS indicators increased fivefold while accuracy reached 77%. This type of indicator estimates the likelihood of companies beating the Wall Street consensus for earnings per share. The team decided to significantly expand this category after token holder CalvinXTZ successfully used these indicators to trade stock options to generate 84% gains in 11 weeks.
For weekly support and resistance level indicators, accuracy is in the 74–80.52% range, depending on the type of indicator. In the spreadsheet, an indicator is considered correct if at least one direction (minimum or maximum price) for the week has been estimated correctly. Let’s turn to indicator interpretation and see why this is appropriate.
Yuri Lobyntsev, Cindicator co-founder & CTO, commented:
“Complex systems like Hybrid Intelligence require time to mature and reach stable performance. Today we can see that our joint efforts have paid off: we’ve validated the wisdom of the crowd theory and managed to build a sustainable asset management platform. We are continuing to refine our mathematical models and execution technology.”
Indicators: interpretations and use cases
Hybrid Intelligence indicators provide a numerical data point for events that are otherwise qualitatively estimated. A CND indicator replaces a gut feeling with a number that can be easily weaved into discretionary trading and algorithmic trading with automated execution.
Let’s look at how the two types of indicators could be used.
Binary indicators
This type of indicator is based on a question that asks analysts to estimate the probability of a given event. If the analyst believes that the event is likely, they will give a value in the range of 50–100% and vice versa. After the answers have been collected, Cindicator’s ML algorithm will assign different weightings to individual forecasts based on the analysts’ past performances, current market conditions, and other data.
Here is an example of a binary question and the resulting indicator:
Interpretation: this particular indicator suggests that there is a high probability of BTC hitting the upper target earlier than declining.
Use case examples:
- Trading assets solely based on indicators with a 1-on-1 risk profile – this uses the indicator’s targets to set stop loss and take profit orders (we outlined this strategy in a previous report);
- Trading with leverage – this uses indicators to select assets and decide on entry and exit points at your discretion. CND holder CalvinXTZ has done this using EPS indicators to trade options (note: this is a higher risk strategy, and you can lose your entire deposit);
- Trading predictions markets using HI indicators – our experiment on Augur markets resulted in an 88% profit in 20 weeks.
These are just some examples of strategies: the range of use cases is really only limited by a trader’s imagination (and market realities).
Support & resistance indicators
This type of indicator estimates a resistance (the maximum price) and support (the minimum price) for the week. This indicator replaces gut-based guessing with a specific number and introduces a new data point for algorithmic trading.
When setting short-term stop loss and take profit orders, a trader usually simply looks at the chart and estimates the likely support and resistance based on previous lows, round numbers, etc. This process is inherently biased and experienced traders try to minimise their own biases while exploiting the cognitive errors of other participants.
Hybrid Intelligence reduces overall bias by collecting hundreds or thousands of forecasts from a large group of incentivised analysts. Cindicator’s ML algorithms then assign different weightings to individual forecasts. The resulting indicator gives two numbers.
Interpretation: this indicator suggests minimum and maximum prices that market participants can expect during the week.
Use case examples:
- Trading reversion style using S&R levels – this requires backtesting to identify specific assets to calculate the trading range (for example, it could be the case that one asset doesn’t break out of the range by more than 15% with high confidence) and to set stop loss and take profit orders accordingly;
- Trading momentum – if the Hybrid Intelligence support or resistance coincides with an important level (e.g. a monthly/quarterly/annual local low or high) and is broken, this could suggest that many traders expect a significant price move, so prices from the indicators could be used in discretionary or algorithmic momentum trading (a backtest is required as well);
- Identifying general motion of ETH or BTC – some token holders who trade altcoins use S&R indicators to gauge the overall expected direction for the major asset against which they have open positions.
- Additional data point for investment – Here is how one token holder puts it: “I first research the fundamentals of crypto projects then for those I have retained I research upcoming news. I further evaluate the charts for these projects using basic technical analysis and it is at this step that I compare my analysis with Cindicator’s Support and Resistance indicators and decide whether I will take action with an investment.”
A word of caution: there is no magic crystal ball
You should remember that any indicator is simply a tool. It could be possible to trade profitably solely on CND indicators, but in our opinion the greater potential lies in a hybrid approach that either:
- Combines the skills of an individual trader and Hybrid Intelligence, e.g. CalvinXTZ or Cindicator team experiments on Augur, or;
- Uses Hybrid Intelligence as additional data input for algorithmic quantitative strategies, as in the Cindiсator Capital approach that generated 23.6% in BTC or 190% in USD in 2019.
Since long-term accuracy is in the 60s, not every indicator will be correct. It is important to spread out risks over multiple trades and remember that every trade has fees.
To create your own optimal strategy, research the historical indicators and select the ones that are most useful to you. Our broad range of indicators has a place in anybody’s strategy, from retail traders who trade altcoins to large portfolio managers or options sellers. We are also testing an API and if you would like to experiment with adding unparsed indicator data sets to your strategy, please fill out the form below.
What’s next
Over the past two years, the accuracy of Cindicator’s analytical products has reached a high and sustainable level. The multi-year track record once again validates Hybrid Intelligence as a valuable tool for predictive analytics.
Mike Brusov, Cindicator co-founder & CEO, said:
“The stable performance opens up nearly unlimited opportunities for scaling the platform and the whole Cindicator ecosystem. While some metrics could be further improved, Hybrid Intelligence now consistently provides value. We believe this approach could be applied to a vast universe of financial markets – digital assets, equities, commodities, etc. – and beyond.
What’s next? Superforecasters, Wisdom of Crowd of Crowds, Ecosystem Intelligence, tokenized funds, crypto robo-advising….or DAO - stay tuned and fasten your belts in 2020.”
The wind rises, we must try to live. And we’ll succeed together.”
Thank you for being a valuable part of the Cindicator ecosystem. More news is coming soon, this is just the beginning!