By Mahavir Bhattacharya
Welcome to the second a part of this two-part weblog sequence on the bias-variance tradeoff and its software to buying and selling in monetary markets.
Within the first half, we tried to develop an instinct for the bias-variance decomposition. On this half, we’ll lengthen our learnings from the primary half and develop a buying and selling technique.
Conditions
When you have some primary data of Python and ML, you need to have the ability to learn and comprehend the article. These are some pre-requisites:
https://weblog.quantinsti.com/bias-variance-machine-learning-trading/Linear algebra (primary to intermediate)Python programming (primary to intermediate)Machine studying (working data of regression and regressor fashions)Time sequence evaluation (primary to intermediate)Expertise in working with market information and creating, backtesting, and evaluating buying and selling methods
Additionally, I’ve added some hyperlinks for additional studying at related locations all through the weblog.
If you happen to’re new to Python or want a refresher on it, you can begin with Fundamentals of Python Programming after which transfer to Python for Buying and selling: Primary on Quantra for trading-specific purposes.
To familiarize your self with machine studying, and with the idea of linear regression, you’ll be able to undergo Machine Studying for Buying and selling and Predicting Inventory Costs Utilizing Regression.
As a result of the article additionally covers time sequence transformations and stationarity, you’ll be able to familiarize your self with Time Sequence Evaluation. Information of dealing with monetary market information and sensible expertise in technique creation, backtesting, and analysis will assist you apply the article’s learnings to your methods.
On this weblog, we’ll cowl the entire pipeline for utilizing machine studying to construct and backtest a buying and selling technique whereas utilising the bias-variance decomposition to pick out the suitable prediction mannequin. So, right here goes…
The circulate of this text is as follows:
As a ritual, step one is to import the required libraries.
Importing Libraries
If you happen to don’t have any of those put in, a ‘!pip set up’ command ought to do the trick (for those who don’t wish to depart the Jupyter Pocket book setting, or if you wish to work on Google Colab).
Downloading Information
Subsequent, we outline a operate for downloading the information. We’ll use the yfinance API right here.
Discover the argument ‘multi_level_index’. Lately (I’m penning this in April 2025), there have been some modifications within the yfinance API. When downloading worth degree and quantity information for any safety via the required API, the ticker identify of the safety will get added as a heading.
It seems to be like this when downloaded:
For folks (like me!) who’re accustomed to not seeing this further degree of heading, eradicating it whereas downloading the information is a good suggestion. So we set the ‘multi_level_index’ argument to ‘False’.
Defining Technical Indicators as Predictor Variables
Subsequent, since we’re utilizing machine studying to construct a buying and selling technique, we should embrace some options (generally referred to as predictor variables) on which we practice the machine studying mannequin. Utilizing technical indicators as predictor variables is a good suggestion when buying and selling within the monetary markets. Let’s do it now.
Finally, we’ll see the record of indicators after we name this operate on the asset dataframe.
Defining the Goal Variable
The following chronological step is to outline the goal variable/s. In our case, we’ll outline a single goal variable, the close-to-close 5-day % return. Let’s see what this implies. Suppose immediately is a Monday, and there are not any market holidays, barring the weekends, this week. Contemplate the % change in tomorrow’s (Tuesday’s) closing worth over immediately’s closing worth, which might be a close-to-close 1-day % return. At Wednesday’s shut, it could be the 2-day % return, and so forth, until the next Monday, when it could be the 5-day % return. Right here’s the Python implementation for a similar:
Why can we use the shift(-5) right here? Suppose the 5-day % return based mostly on the closing worth of the next Monday over immediately’s closing worth is 1.2%. By utilizing shift(-5), we’re putting this worth of 1.2% within the row for immediately’s OHLC worth ranges, quantity, and different technical indicators. Thus, after we feed the information to the ML mannequin for coaching, it learns by contemplating the technical indicators as predictors and the worth of 1.2% in the identical row because the goal variable.
Stroll Ahead Optimisation with PCA and VIF
One important consideration whereas coaching ML fashions is to make sure that they show sturdy generalization. Which means that the mannequin ought to have the ability to extrapolate its efficiency on the coaching dataset (generally referred to as in-sample information) to the take a look at dataset (generally referred to as out-of-sample information), and its good (or in any other case) efficiency must be attributed primarily to the inherent nature of the information and the mannequin, quite than to probability.
One strategy in the direction of that is combinatorial purged cross-validation with embargoing. You may learn this to study extra.
One other strategy is walk-forward optimisation, which we are going to use (learn extra: 1 2).
One other important consideration whereas constructing an ML pipeline is function extraction. In our case, the entire predictors now we have is 21. We have to extract a very powerful ones from these, and for this, we are going to use Principal Part Evaluation and the Variance Inflation Issue. The previous extracts the highest 4 (a price that I selected to work with; you’ll be able to change it and see how the backtest modifications) combos of options that specify essentially the most variance inside the dataset, whereas the latter addresses mutual data, often known as multicollinearity.
Right here’s the Python implementation of constructing a operate that does the above:
Buying and selling Technique Formulation, Backtesting, and Analysis
We now come to the meaty half: the technique formulation. Listed below are the technique outlines:
Preliminary capital: ₹10,000.
Capital to be deployed per commerce: 20% of preliminary capital (₹2,000 in our case).
Lengthy situation: when the 5-day close-to-close % return prediction is optimistic.
Brief situation: when the 5-day close-to-close % return prediction is adverse.
Entry level: open of day (N+1). Thus, if immediately is a Monday, and the prediction for the 5-day close-to-close % returns is optimistic immediately, I’ll go lengthy at Tuesday’s open, else I’ll go brief at Tuesday’s open.
Exit level: shut of day (N+5). Thus, after I get a optimistic (adverse) prediction immediately and go lengthy (brief) throughout Tuesday’s open, I’ll sq. off on the closing worth of the next Monday (offered there are not any market holidays in between).
Capital compounding: no. Which means that our earnings (losses) from each commerce should not getting added (subtracted) to (from) the tradable capital, which stays mounted at ₹10,000.
Right here’s the Python code for this technique:
Subsequent, we outline the features to guage the Sharpe ratio and most drawdowns of the technique and a buy-and-hold strategy.
Calling the Features Outlined Beforehand
Now, we start calling among the features talked about above.
We’ll begin with downloading the information utilizing the yfinance API. The ticker and interval are user-driven. When operating this code, you’ll be prompted to enter the identical. I selected to work with the 10-year day by day information of the NIFTY-50, the broad market index based mostly on the Nationwide Inventory Change (NSE) of India. You may select a smaller timeframe; the longer the timeframe, the longer it’s going to take for the next codes to run. After downloading the information, we’ll create the technical indicators by calling the ‘create_technical_indicators’ operate we outlined beforehand.
Right here’s the output of the above code:
Enter a legitimate yfinance API ticker: ^NSEI
Enter the variety of years for downloading information (e.g., 1y, 2y, 5y, 10y): 10y
YF.obtain() has modified argument auto_adjust default to True
[*********************100%***********************] 1 of 1 accomplished
Subsequent, we align the information:
Let’s test the 2 dataframes ‘indicators’ and ‘data_merged’.
RangeIndex: 2443 entries, 0 to 2442
Information columns (whole 21 columns):
# Column Non-Null Rely Dtype
— —— ————– —–
0 sma_5 2443 non-null float64
1 sma_10 2443 non-null float64
2 ema_5 2443 non-null float64
3 ema_10 2443 non-null float64
4 momentum_5 2443 non-null float64
5 momentum_10 2443 non-null float64
6 roc_5 2443 non-null float64
7 roc_10 2443 non-null float64
8 std_5 2443 non-null float64
9 std_10 2443 non-null float64
10 rsi_14 2443 non-null float64
11 vwap 2443 non-null float64
12 obv 2443 non-null int64
13 adx_14 2443 non-null float64
14 atr_14 2443 non-null float64
15 bollinger_upper 2443 non-null float64
16 bollinger_lower 2443 non-null float64
17 macd 2443 non-null float64
18 cci_20 2443 non-null float64
19 williams_r 2443 non-null float64
20 stochastic_k 2443 non-null float64
dtypes: float64(20), int64(1)
reminiscence utilization: 400.9 KB
Index: 2438 entries, 0 to 2437
Information columns (whole 28 columns):
# Column Non-Null Rely Dtype
— —— ————– —–
0 Date 2438 non-null datetime64[ns]
1 Shut 2438 non-null float64
2 Excessive 2438 non-null float64
3 Low 2438 non-null float64
4 Open 2438 non-null float64
5 Quantity 2438 non-null int64
6 sma_5 2438 non-null float64
7 sma_10 2438 non-null float64
8 ema_5 2438 non-null float64
9 ema_10 2438 non-null float64
10 momentum_5 2438 non-null float64
11 momentum_10 2438 non-null float64
12 roc_5 2438 non-null float64
13 roc_10 2438 non-null float64
14 std_5 2438 non-null float64
15 std_10 2438 non-null float64
16 rsi_14 2438 non-null float64
17 vwap 2438 non-null float64
18 obv 2438 non-null int64
19 adx_14 2438 non-null float64
20 atr_14 2438 non-null float64
21 bollinger_upper 2438 non-null float64
22 bollinger_lower 2438 non-null float64
23 macd 2438 non-null float64
24 cci_20 2438 non-null float64
25 williams_r 2438 non-null float64
26 stochastic_k 2438 non-null float64
27 Goal 2438 non-null float64
dtypes: datetime64[ns](1), float64(25), int64(2)
reminiscence utilization: 552.4 KB
The dataframe ‘indicators’ incorporates all 21 technical indicators talked about earlier.
Bias-Variance Decomposition
Now, the first goal of this weblog is to exhibit how the bias-variance decomposition can help in creating an ML-based buying and selling technique. In fact, we aren’t simply limiting ourselves to it; we’re additionally studying the entire pipeline of making and backtesting an ML-based technique with robustness. However let’s discuss in regards to the bias-variance decomposition now.
We start by defining six totally different regression fashions:
You may add extra or subtract a pair from the above record. The extra regressor fashions there are, the longer the next codes will take to run. Lowering the variety of estimators within the related fashions will even end in quicker execution of the next codes.
In case you’re questioning why I selected regressor fashions, it’s as a result of the character of our goal variable is steady, not discrete. Though our buying and selling technique is predicated on the path of the prediction (bullish or bearish), we’re coaching the mannequin to foretell the 5-day return, a steady random variable, quite than the market motion, which is a categorical variable.
After defining the fashions, we outline a operate for the bias-variance decomposition:
You may lower the worth of num_rounds to, say, 10, to make the next code run quicker. Nonetheless, a better worth offers a extra sturdy estimate.
This can be a good repository to search for the above code:
https://rasbt.github.io/mlxtend/user_guide/consider/bias_variance_decomp/
Lastly, we run the bias-variance decomposition:
The output of this code is:
Bias-Variance Decomposition for All Fashions:
Complete Error Bias Variance Irreducible Error
LinearRegression 0.000773 0.000749 0.000024 -2.270048e-19
Ridge 0.000763 0.000743 0.000021 1.016440e-19
DecisionTree 0.000953 0.000585 0.000368 -2.710505e-19
Bagging 0.000605 0.000580 0.000025 7.792703e-20
RandomForest 0.000605 0.000580 0.000025 1.287490e-19
GradientBoosting 0.000536 0.000459 0.000077 9.486769e-20
Let’s analyse the above desk. We’ll want to decide on a mannequin that balances bias and variance, that means it neither underfits nor overfits. The choice tree regressor finest balances bias and variance amongst all six fashions.
Nonetheless, its whole error is the very best. Bagging and RandomForest show related whole errors. GradientBoosting shows not simply the bottom whole error but in addition a better diploma of variance in comparison with Bagging and RandomForest; thus, its skill to generalise to unseen information must be higher than the opposite two, since it could seize extra advanced patterns..
You may be compelled to suppose that with such proximity of values, such in-depth evaluation isn’t apt owing to a excessive noise-to-signal ratio. Nonetheless, since we’re operating 100 rounds of the bias-variance decomposition, we will be assured within the noise mitigation that outcomes.
Lengthy story minimize brief, we’ll select to coach the GradientBoosting regressor, and use it to foretell the goal variable. You may, after all, change the mannequin and see how the technique performs below the brand new mannequin. Please word that we’re treating the ML fashions as black bins right here, as exploring their underlying mechanisms is outdoors the scope of this weblog. Nonetheless, when utilizing ML fashions for any use case, we should always all the time pay attention to their inside workings and select accordingly.
Having mentioned all of the above, is there a approach of lowering the errors of a number of of the above regressor fashions? Sure, and it’s not a mere method, however an integral a part of working with time sequence. Let’s talk about this.
Stationarising the Inputs
We’re working with time sequence information (learn extra), and when performing monetary modeling duties, we have to test for stationarity (learn extra). In our case, we should always test our enter variables (the predictors) for stationarity. Let’s test the predictor variables for stationarity and apply differencing to the required predictors (learn extra).
Right here’s the code:
Right here’s a snapshot of the output of the above code:
The above output signifies that 13 predictor variables don’t require stationarisation, whereas 8 do. Let’s stationarise them.
Let’s confirm whether or not the stationarising bought accomplished as anticipated or not:
Yup, accomplished!
Let’s align the information once more:
Let’s test the bias-variance decomposition of the fashions with the stationarised predictors:
Right here’s the output:
Bias-Variance Decomposition for All Fashions with Stationarised Predictors:
Complete Error Bias Variance Irreducible Error
LinearRegression 0.000384 0.000369 0.000015 5.421011e-20
Ridge 0.000386 0.000373 0.000013 -3.726945e-20
DecisionTree 0.000888 0.000341 0.000546 2.168404e-19
Bagging 0.000362 0.000338 0.000024 -1.151965e-19
RandomForest 0.000363 0.000338 0.000024 7.453890e-20
GradientBoosting 0.000358 0.000324 0.000034 -3.388132e-20
There you go. Simply by following Time Sequence 101, we may scale back the errors of all of the fashions. For a similar cause that we mentioned earlier, we’ll select to run the prediction and backtesting utilizing the GradientBoosting regressor.
Working a Prediction utilizing the Chosen Mannequin
Subsequent, we run a walk-forward prediction utilizing the chosen mannequin:
Now, we create a dataframe, ‘final_data’, that incorporates solely the open costs, shut costs, precise/realised 5-day returns, and 5-day returns predicted by the mannequin. We want the open and shut costs for getting into and exiting trades, and the expected 5-day returns, to find out the path wherein we take trades. We then name the ‘backtest_strategy’ operate on this dataframe.
Checking the Commerce Logs
The dataframe ‘trades_df_differenced’ incorporates the commerce logs.
We’ll convert the decimals of the values within the dataframe for higher visibility:
Let’s test the dataframe ‘trades_df_differenced’ now:
Right here’s a snapshot of the output of this code:
From the desk above, it’s obvious that we take a brand new commerce day by day and deploy 20% of our tradeable capital on every commerce.
Fairness Curves, Sharpe, Drawdown, Hit Ratio, Returns Distribution, Common Returns per Commerce, and CAGR
Let’s calculate the fairness for the technique and the buy-and-hold strategy:
Subsequent, we calculate the Sharpe and the utmost drawdowns:
The above code requires you to enter the risk-free fee of your alternative. It’s sometimes the federal government treasury yield. You may look it up on-line on your geography. I selected to work with a price of 6.6:
Enter the risk-free fee (e.g., for five.3%, enter solely 5.3): 6.6
Now, we’ll reindex the dataframes to a datetime index.
We’ll plot the fairness curves subsequent:
That is how the technique and buy-and-hold fairness curves look when plotted on the identical chart:
The technique fairness and the underlying transfer virtually in tandem, with the technique underperforming earlier than the COVID-19 pandemic and outperforming afterward. Towards the top, we’ll talk about some real looking concerns about this relative efficiency.
Let’s take a look on the drawdowns of the technique and the buy-and-hold strategy:
Let’s check out the Sharpe ratios and the utmost drawdown by calling the respective features that we outlined earlier:
Output:
Sharpe Ratio (Technique with Stationarised Predictors): 0.89
Sharpe Ratio (Purchase & Maintain): 0.42
Max Drawdown (Technique with Stationarised Predictors): -11.28%
Max Drawdown (Purchase & Maintain): -38.44%
Right here’s the hit ratio:
Hit Ratio of Technique with Stationarised Predictors: 54.09%
That is how the distribution of the technique returns seems to be like:
Lastly, let’s calculate the common earnings (losses) per successful (shedding) commerce:
Common Revenue for Worthwhile Trades with Stationarised Predictors: 0.0171
Common Loss Loss-Making Trades with Stationarised Predictors: -0.0146
Primarily based on the above commerce metrics, we revenue extra on common in every commerce than we lose. Additionally, the variety of optimistic trades exceeds the variety of adverse trades. Due to this fact, our technique is protected on each fronts. The utmost drawdown of the technique is proscribed to 10.48%.
The explanation: The holding interval for any commerce is 5 days, utilizing solely 20% of our obtainable capital per commerce. This additionally reduces the upside potential per commerce. Nonetheless, for the reason that common revenue per worthwhile commerce is larger than the common loss per loss-making commerce and the variety of worthwhile trades is larger than the variety of loss-making trades, the probabilities of capturing extra upsides are larger than these of capturing extra downsides.
Let’s calculate the compounded annual development fee (CAGR):
CAGR (Purchase & Maintain): 13.0078%
CAGR (Technique with Stationarised Predictors): 13.3382%
Lastly, we’ll consider the regressor mannequin’s accuracy, precision, recall, and f1 scores (learn extra).
Confusion Matrix (Stationarised Predictors):
[[387 508]
[453 834]]
Accuracy (Stationarised Predictors): 0.5596
Recall (Stationarised Predictors): 0.6480
Precision (Stationarised Predictors): 0.6215
F1-Rating (Stationarised Predictors): 0.6345
Some Real looking Issues
Our technique outperformed the underlying index throughout the post-COVID-19 crash interval and marginally outperformed the general market. Nonetheless, in case you are considering of utilizing the skeleton of this technique to generate alphas, you’ll have to peel off some assumptions and bear in mind some real looking concerns:
Transaction Prices: We enter and exit trades day by day, as we noticed earlier. This incurs transaction prices.
Asset Choice: We backtested utilizing the broad market index, which isn’t instantly tradable. We’ll want to decide on ETFs or derivatives with this index because the underlying.
Slippages: We enter our trades on the market’s opening and exit at its shut. Buying and selling exercise will be excessive throughout these durations, and we could encounter appreciable slippages.
Availability of Partially Tradable Securities: Our backtest implicitly assumes the provision of fractional property. For instance, if our capital is ₹2,000 and the entry worth is ₹20,000, we’ll have the ability to purchase or promote 0.1 items of the underlying, ignoring all different prices.
Taxes: Since we’re getting into and exiting trades inside very brief time frames, other than transaction prices, we might incur a major quantity of short-term capital beneficial properties tax (STCG) on the earnings earned. This, after all, would rely in your native rules.
Danger Administration: Within the backtest, we omitted stop-losses and take-profits. You’re inspired to incorporate them and tell us your findings on how the technique’s efficiency will get modified.
Occasion-driven Backtesting: The backtesting we carried out above is vectorized. Nonetheless, in actual life, tomorrow comes solely after immediately, and we should think about this when performing a backtest. You may discover the Blueshift at https://blueshift.quantinsti.com/ and take a look at backtesting the above technique utilizing an event-driven strategy (learn extra). An event-driven backtest would additionally account for slippage, transaction prices, implementation shortfalls, and danger administration.
Technique Efficiency: The hit ratio of the technique and the mannequin’s accuracy are roughly 54% and 56%, respectively. These values are marginally higher than these of a coin toss. You need to do that technique with different asset lessons and solely choose these property on which these values are no less than 60% (or larger for those who wanna be extra conservative). Solely after that ought to you carry out an event-driven backtesting utilizing this technique define.
A Word on the Downloadable Python Pocket book
The downloadable pocket book includes backtesting the technique and evaluating its efficiency and the mannequin’s efficiency parameters in a state of affairs the place the predictors should not stationarised and after stationarising them (as we noticed above). Within the former, the technique considerably outperforms the underlying mannequin, and the mannequin shows larger accuracy in its predictions regardless of its larger errors displayed throughout the bias-variance decomposition. Thus, a well-performing mannequin needn’t essentially translate into a very good buying and selling technique, and vice versa.
The Sharpe of the technique with out the predictors stationarised is 2.56, and the CAGR is nearly 27% (versus 0.94 and 14% respectively when the predictors are stationarised). Since we used GradientBoosting, a tree-based mannequin that does not essentially want the predictor variables to be stationarised, we will work with out stationarising the predictors and reap the advantages of the mannequin’s excessive efficiency with non-stationarised predictors.
Word that operating the pocket book will take a while. Additionally, the performances you get hold of will differ a bit from what I’ve proven all through the article.
There’s no ‘Good’ in Goodbye…
…but, I’ll should say so now 🙂. Check out the backtest with totally different property by altering among the parameters talked about within the weblog, and tell us your findings. Additionally, as we all the time say, since we aren’t a registered funding advisory, any technique demonstrated as a part of our content material is for demonstrative, instructional, and informational functions solely, and shouldn’t be construed as buying and selling or funding recommendation. Nonetheless, for those who’re capable of incorporate all of the aforementioned real looking elements, extensively backtest and ahead take a look at the technique (with or with out some tweaks), generate important alpha, and make substantial returns by deploying it within the markets, do share the excellent news with us as a remark under. We’ll be completely satisfied on your success 🙂. Till subsequent time…
Credit
José Carlos Gonzáles Tanaka and Vivek Krishnamoorthy, thanks on your meticulous suggestions; it helped form this text!Chainika Thakar, thanks for rendering this and making it obtainable to the world!
Subsequent Steps
After going via the above, you’ll be able to observe just a few structured studying paths if you wish to broaden and/or deepen your understanding of buying and selling mannequin efficiency, ML technique improvement, and backtesting workflows.
To grasp every part of this technique — from Python and PCA to stationarity and backtesting — discover topic-specific Quantra programs like:
For these aiming to consolidate all of this information right into a structured, mentor-led format, the Govt Programme in Algorithmic Buying and selling (EPAT) gives a really perfect subsequent step. EPAT covers every little thing from Python and statistics to machine studying, time sequence modeling, backtesting, and efficiency metrics analysis — equipping you to construct and deploy sturdy, data-driven methods at scale.
File within the obtain:
Bias Variance Decomposition – Python pocket book
Be at liberty to make modifications to the code as per your consolation.
Login to Obtain
All investments and buying and selling within the inventory market contain danger. Any determination to position trades within the monetary markets, together with buying and selling in inventory or choices or different monetary devices is a private determination that ought to solely be made after thorough analysis, together with a private danger and monetary evaluation and the engagement {of professional} help to the extent you imagine needed. The buying and selling methods or associated data talked about on this article is for informational functions solely.
