FAQ#
Here, we answer common questions about our project, usage, and best practices.
Is there any platform given by Zelta to submit alpha?#
Yes, there is a platform provided by Zelta called Jupyter_untrade for testing, developing, refining, and pushing code.
Platform link: jupyter.untrade.io
What is the team size and nature of work?#
The team size is spread across both onsite and remote locations, consistently managing running alphas and simultaneously working on alpha research.
The nature of work involves creating, deploying, and optimizing quantitative trading strategies on bitcoin that outperform benchmarks.
What is the compensation for this Create/Earn project?#
The Create and Earn project includes a profit-sharing component, the details of which are clearly mentioned and defined in the documentation.
Is the requirement onsite or remote?#
For the Create and Earn project, there is no geographical dependency, and the work is completely remote-based.
What is the deadline for the project?#
There is no deadline for the Create/Earn project. Candidates are expected to work at their own pace. However, participants are encouraged to work briskly as the program involves a three-week front_run monitoring period before official commencement.
What next after alpha is submitted?#
Once the alpha consistently generates profit on a quarterly/semi-annual/yearly basis and completes a full bitcoin cycle (2020-2023), monitoring_frontrun for three weeks will follow, followed by the official start of Day 1_Create/Earn.
How shall we proceed since we don’t have any experience with respect to creating alphas and we are freshers in quantitative trading?#
We encourage you to align with us and start with basic approaches since the project has no deadline. Meanwhile, you can learn and grow with our suggestions, feedback, and your efforts in improving iteratively every day
Can this program also help us in getting Full Time opportunities at ZeltaLab’s Quant Desk?#
Alpha submission component is common in both Create/Earn and Full Time. However for Full Time- there is an additional layer of code optimality test along with culture fit round. Requirements can be both onsite/remote and depend on the candidate’s discretion and will.
What data will participants have access to, and for what time period?#
Participants will have access to BTC/USDT OHCLV data spanning different candlestick time frames from January 1, 2020, to December 31, 2023. Logically categorize your train , val and test set in this.
What is the process of submitting a trading strategy and what happens next?#
Source code of the algorithmic trading strategies developed for the BTC/USDT market, along with documentation to be pushed on ZeltaLabs’ Jupyter_Untrade..
Strategy will be deployed in a live trading environment and kept under monitoring for 3 weeks. Once the price-timestamp coherence is there, we can start your official day 1 of Create/Earn.
During the three-week observation period in a live trading environment, we ensure coherence between back-testing logs and real-time front test orders, assess code functionality, and evaluate the impact of real-life order books on net profitability.
What would be the standard metrics for evaluating the performance of an alpha?#
- Any strategy should outperform the benchmark (buy and hold returns of BTC); otherwise, trading is essence-less. We are considering the following parameters:
Annual return: 50-100% Quarterly return percentage: Exceeding the buy and hold benchmark Win rate: Equal to or greater than 50% Sharpe ratio: Equal to or greater than 3 Max drawdown: Less than or equal to 10%. Max dip (maximum adverse exposure) : Less than or equal to 6-7%.
What are the suggested approaches for strategy development?#
Participants are encouraged to explore statistical analysis and various mathematical approaches to create alpha-producing strategies, balancing computational efficiency with maximized returns.
- Some approaches in Algorithmic Trading that can be employed for this program:-
Machine learning models like regression models, decision trees, random forests, support vector machines etc. Deep learning models, such as recurrent neural networks and convolutional neural networks etc. Reinforcement learning techniques, such as Q-learning or deep reinforcement learning etc. Technical analysis i.e statistical approaches using various indicators or combinations of indicators to generate noise free signals etc. Various miscellaneous approaches like bagging and boosting for accurate prediction , monte carlo methods or stochastic models like Markov chains etc.
Can we open multiple short or long trades at a time or do we need to close the previous trade that we have opened and then open a new one?#
Close the previous trade that we have opened and then open a new one.Therefore we just open one trade at a time and square it off before opening a new one .We would also suggest you to enter once and exit entirely once for the strategy part since pyramiding is not allowed in this.
Could you please elaborate on how benchmark returns are calculated and why it is an important metric for us?#
Benchmark returns are calculated by buying BTC on the first day of the year and holding it until the last day of the year. This simple buy-and-hold strategy serves as a baseline comparison for evaluating the performance of trading strategies over the same time frame.
This comparison helps evaluate the performance of the trading strategy and strategies should try to universally outperform benchmarks irrespective of time periods and sentiments in markets.
Can we use leverage in our strategy to beat the BTC benchmarks?#
Leverage can be used if your maximum drawdowns and maximum dips are meticulously contained in your strategies. However, we recommend producing consistent alphas without leveraging.
Is the use of any external libraries allowed in the code, such as backtrader, backtesting.py, etc.?#
The use of any external library in the submitted code is strictly prohibited, including libraries like backtrader, backtesting.py, etc. We will also assess the implementation of Object-Oriented Programming (OOP) principles in the code, which promote code modularity and maintainability.
How are commissions calculated in Zelta’s in-house backtesting engine?#
Initially we went ahead with official binance charges(0.10%). However we realized that hidden costs like slippages should be included in the backtesting engine(0.15%) for more clearer representation of metrics like profit and loss.
For every trade - you are shelling out 1.5$ dollar (0.15% of 1000 dollars in static approach)
Supposedly you have 2000 trades in entire dataset then = 2000*1.5$ amounting to 3000$ (final outgoing charges including commission + slippages).
Is usage of backtesting libraries of any kind like backtrader , bakctesting.py , vectorbt etc. allowed, for signal generation ?#
Use of any external library in submitted code is not allowed/strictly prohibited such as backtrader , backtesting.py etc.
We suggest implementing Object-Oriented Programming (OOP) principles in the code, which promote code modularity and maintainability.
We expect candidates to use Zelta’s in-house library for signal generation and backtesting.
If there is a signal 1 at timestamp x, will the buy order be executed at the closing price of timestamp x or opening price of timestamp x+1 ?#
If you have given 1 at timestamp x then buy order will be executed at closing of “timestamp x “ or “opening price of timestamp x+1” which is the same.
The data that you have provided is the only data we can use ?#
Participants are not required to use any data other than the data we have provided. Logically categorize your train , validation and test set.
What are the benefits of participating in this program?#
Participants have the opportunity to enhance their skills, gain practical experience, receive mentorship from industry experts, and potentially secure opportunities in Zelta Labs’ Quant Desk or high-frequency trading firms securing your candidature.
Can participants make multiple submissions of strategies?#
Yes, participants can make multiple submissions, and further decisions regarding final submissions will be mutually agreed upon for assessment.
Who owns the intellectual property developed during the program?#
All intellectual property developed during the program belongs to Zelta Labs.
What is the formal agreement process for participation?#
Participants will sign an MOU/NDA to formalize the agreement, ensuring clarity and confidentiality in the partnership.
What are the position sizing approaches to be considered for trades?#
You can use two approaches:-
Using 1000 dollars in every trade. (static)
Opening the first trade with 1000 dollars and then adjusting the profit/loss of that trade in 1000 dollars. Whatever balance is left after adjustment , next trade will be executed with that total balance. It goes on arithmetically like this. (compounding).
Why do we prefer statistical approaches rather than ML-centric approaches ?#
We’ve noticed some common concerns in all ML models, such as challenges in generalizing capabilities on unseen data, suboptimal performance on hidden data patches, failing of execution in live trading environments due to non-explainability of models. We acknowledge the importance of periodic retraining for ML models to enhance generalization capabilities.
Understanding why a model produces specific signals and its behavior is crucial for effective strategy development- Explainable AI/XML here plays a crucial role in addressing the challenges thrown by black box.
Under Create/Earn , do we need to prepare a comprehensive documentation of our alpha strategy also?#
Yes , document needs to be submitted post alpha-submission, describing the underlying principles, indicators, algorithms, or models used in your strategy. Explain how your strategy identifies trading opportunities and makes decisions.
Incorporate charts, graphs, and visualizations to illustrate key concepts, trends, and performance metrics effectively.
Summarize the main findings, lessons learned, and takeaways from developing and testing your alpha strategy.
Do we have to consider developing a long-only, short-only, or long-short trading strategy?#
We personally suggest to go with both long and short positions in a strategy , allowing traders to take advantage of both upward and downward price movements in the market. This can help mitigate risks and potentially enhance returns by diversifying the portfolio and exploit opportunities in different market environments.
Can we choose to work on any timeframe/frequency?#
Yes, you can choose your own timeframe. We have shared frequency from 3 minutes to 1D therefore you are at your own liberty to choose any frequency and build your strategies.
If we’re using an ML model with a 3:1 training to testing split (or any other ratio), then do we have to submit logs/signals for all the data or just the test data?#
Just the signals for the test split.
Are we expected to do intraday trading, swing trading, or regime trading?#
The participants are expected to devise various strategies catering to intraday, swing, or regime trading. Ensure that intraday strategies beat quarterly benchmark returns on BTC, while swing opportunities should exceed semi-annual benchmark returns.
Will participants receive mentoring during the program?#
Yes, participants will receive mentoring from Zelta’s Quant Desk members to guide them in creating their own strategies, fostering personal and intellectual growth.
Can we get formulas for calculating PnL and other important metrics?#
Formulas are widely available on the internet. We have provided formulas for calculating PnLs of all logs and then eventually arriving at a cumulative PnL. Additionally, you can use our library, which has a backtesting engine integrated into it for quick self-assessment of metrics in your approaches.
Once it is generating alpha consistently quarterly/semi-annually/yearly and on a complete bitcoin cycle of 4 years , we can go ahead with monitoring_frontrun for 3 weeks followed by official start of Day 1_Create/Earn.
What is the deadline for the project?#
There is no strict deadline for the Create/Earn initiative. However, participants are encouraged to work at a brisk pace as the program involves a three-week front_run monitoring period before official commencement.