Skip to Main Content

03/05/2024 |

A Revolutionary Approach To Strategy Optimization: The Strategy Variance Explorer

This is an image that promotes the tax time sale in which you can save up to 73% on TrendSpider, this weekend only.

Understanding and analyzing variance across trading strategies is fundamental for honing your tactics in the markets, and here at TrendSpider, we are thrilled to introduce a brand new tool that makes this possible on the platform; The Strategy Variance Explorer. With it, you can now:

  • Assess the robustness of your strategies
  • Identify the optimal parameters utilized within your strategies
  • Compare strategy performance
  • Enhance your risk management

The Strategy Variance Explorer revolutionizes strategy testing by offering a comprehensive toolset to explore strategy variations across different time frames, symbols, and methodologies. This upgrade empowers you with unparalleled data access and analysis capabilities, enabling you to optimize your strategies for maximum profitability.

This is the header image for the strategy variance explorer blog post.

A World Of Data At Your Fingertips

The Strategy Variance Explorer tool opens up a world of possibilities and truly an unbelievable amount of data. With just a few clicks, you can now:

  • Test one strategy on multiple time frames or symbols to identify the most promising options.
  • Compare how metrics change depending on symbols, time frames, or strategies.
  • Determine the strategy of best fit by testing multiple strategies on the same chart, symbol, and time frame.
  • Experiment with various combinations to gain deeper insights into your trading strategies.

How The Tool Is Comprised

The Strategy Variance Explorer widget is your gateway to this powerful tool. It can be found in the top right corner of your Strategy Tester and consists of three main parts:

This is an image of the Variance Tester button found in the top right corner of the Strategy Tester.

  1. Configuration Panel: Here, you set up your variance test by selecting the strategies, time frames, depth, and symbols to test. At this time, a total of 30 variants can be tested at once.
  2. Variant Table: This spreadsheet displays tabular metrics for each backtest variant, allowing you to compare their performance easily. The brightest green highlight represents the best data point.
  3. Bubble Charts: Three bubble charts offer visual representations of your variant table data. You have the flexibility to customize which metrics represent the horizontal axis, vertical axis, and bubble size. The bubbles are color-coded on a gradient from orange to green. The best combinations of X and Y metrics show as green bubbles, the worst as orange, and the rest with colors in between. While default metrics are set, you can easily change them to suit your analysis.

This is an image of the full variance explorer panel.

Ways To Use The Variance Tester

Multi-Time Frame Testing

Multi-time frame testing allows you to test one strategy on the same asset across time frames. This style of testing enables you to hone in on the time frames that are most suited for your strategies, as well as those that should be avoided.

In the below example, we’ve tested an 8/21 EMA bullish cross on the QQQ ETF across 10 different time frames ranging between the 10-minute and the Daily. This strategy buys when the 8 EMA crosses up on the 21 EMA and sells when the 8 EMA crosses down on the 21 EMA.

This is an image of the variance explorer results of the strategy defined above. An explanation of this image can be found in the bullet points below.

Insights From This Test

  • Configuration Panel: The daily time frame had the highest percentage return, the 2hr took the most positions, and the 4hr and daily were tied for the highest win rate percentage.
  • R/R / Win Rate Bubble Chart: The daily time frame had the highest R/R value relative to the win rate, the 120m has the lowest, and the net return of all time frames contained within the red-shaded area can be expected to go to zero over time.
  • Avg. Return / St. Dev Return Bubble Chart: The daily time frame has the highest average return and the 10m has the lowest. In the best-case scenario, traders might want to look for the highest average return with the lowest deviation between returns.
  • Max DD / Max DD (asset) Bubble Chart: The 10m time frame draws down the least, but relative to the underlying asset, the 15m is the best performer in this category. The 4hr has the highest max drawdown.

Multi-Symbol Testing

Multi-symbol testing allows you to identify which asset your strategy works best on. If testing across assets that should have similar metrics, you can also discover anomalies that can be used to make more informed trading decisions.

In the below example, we’ve tested the same 8/21 EMA bullish cross, but this time, on the Magnificent 7 names; AAPL, NVDA, MSFT, META, TSLA, GOOG, and AMZN. We also focused on a single time frame, the daily.

In this image, we see the strategy variance results of the 8/21 ema strategy tested against the names in the mag seven. The insights from this test are listed below this image.

Insights From This Test

  • Configuration Panel: TSLA was the best performer of the group and the only name that beat buy and hold over the same period. AAPL had the lowest max drawdown, while NVDA and MSFT were tied for most positions taken.
  • Positions / Wins Bubble Chart: MSFT and AAPL had the most winning positions, but TSLA had the highest average return per position.
  • Loss Streak / Losses Bubble Chart: META was the worst performer here, with the highest losing streak to total loss ratio.
  • Win Streak / Wins Bubble Chart: AAPL was the best performer here, with the highest win streak to total win ratio.

Multi-Strategy Testing

Multi-strategy testing allows you to compare the results of multiple strategies to find out which strategy works the best on the asset you’re testing.

In the below example, we’ve tested six different 15m strategies on TSLA. There is a mix of bullish and bearish strategies utilizing a range of indicator and price action setups, including Bollinger Bands, RSI, Fair Value Gaps, Supertrend, Opening Range Breakouts, and more.

This is an image of the multi-strategy variance test for TSLA. The results of this test are detailed in the text below.

Insights From This Test

  • Configuration Panel: The ‘Momentum Scalper’ strategy was the best performer of the group with the highest risk-to-reward ratio, as well. The ‘ORB long’ took the most positions and had the highest win rate, while the ‘RSI w/ BBand Long’ strategy had the highest drawdown.
  • Average Win / Max DD Bubble Chart: The ‘Momentum Scalper’ strategy has the best ratio, while the ‘RVOL reversal’ strategy has the worst.
  • R. vol / R. vol (asset) Bubble Chart: All six strategies have a lower realized volatility than buy and hold, but the ‘RVOL reversal’ strategy has the lowest while the ‘RSI w/ BBand’ strategy has the highest.
  • Trades (day) / Trades (month) Bubble Chart: The ‘OBR long’ strategy trades the most often, while the ‘RVOL reversal’ and ‘RSI w/ BBand long’ strategies trade the least often.

Combination Testing

With combination testing, you can utilize up to 30 variants to test a set of names, time frames, and strategies at once. This is the most complex style of variance test that can be run.

In the below example, we’ve tested 3 bullish moving average cross strategies (8/21 EMA, 5/20 SMA, 50/200 SMA) on 3 symbols (QQQ, SPY, IWM) across 3 time frames (15m, 1hr, 1d).

In this image, we see the variance data for all of the strategies listed above. The highlights from this data are in the text below.

Insights From This Test

  • Configuration Panel: The golden cross on the QQQ daily time frame was the best-performing strategy of the group. The worst was the golden cross on the IWM 60m time frame. The 5/20 SMA cross on the IWM daily had the highest drawdown, while the golden cross on the QQQ daily had the highest average return and risk to reward.
  • R/R / Win Rate Bubble Chart: Again here, we see the daily golden cross on the QQQ had the best positioning, while the daily golden cross on the IWM had the worst.
  • Avg. Return / St. Dev Return Bubble Chart: The daily golden cross on the QQQ had the highest average return but also the highest deviation of returns, as well. The SPY daily golden cross has just a slightly lower average return, but also, a slightly lower deviation of returns.
  • Max DD / Max DD (asset) Bubble Chart: Interestingly, all three 15m golden cross strategy on SPY had the lowest max drawdown, while the daily golden cross on the IWM had the highest.

Current Limitations

While the Strategy Variance Explorer offers groundbreaking capabilities, it’s essential to be aware of its current limitations:

  • The tool is in beta, so expect some hiccups along the way.
  • You’re limited to testing 30 variants at a time.
  • Customization data is not preserved between sessions.

Despite these limitations, the Strategy Variance Explorer represents a significant leap forward in backtesting methodologies, empowering traders to make more informed decisions based on comprehensive data analysis.

Strats To Get You Started

Need a few sample strategies to get you started on Variance Testing? We’ve got you covered. Click the below links to download these moving average strategies to your account today.

"$QQQ Golden Cross (Daily)" strategy by TrendSpider
charts.trendspider.com
"$QQQ Golden Cross (Daily)" strategy by TrendSpider
"TSLA Momentum Scalper Long" strategy by TrendSpider
charts.trendspider.com
"TSLA Momentum Scalper Long" strategy by TrendSpider

In conclusion, if you’re serious about refining your trading strategies and maximizing your profits, the Strategy Variance Explorer is a must-have in your toolkit. To learn more, have a read through the Developer Documentation. Try it out today and unlock the full potential of your trading endeavors!