Simpler models prove more stable in crypto trading test

Simpler statistical models delivered more stable results than a deep-learning model in a cryptocurrency trading experiment conducted by three graduates of WorldQuant University’s Financial Engineering programme. Sydney Anuyah, Ogenna Ehiemere and Oluwarotimi Ogundele tested different modelling approaches using Bitcoin and Ethereum market data as part of their master’s capstone project. The researchers compared statistical methods, including Vector Autoregression (VAR), with Long Short-Term Memory (LSTM), a deep-learning model used for analysing sequential data. According to Anuyah, the LSTM recorded strong results in some of the experiments but its performance varied, while the statistical approaches produced more consistent results and were easier to interpret. “The most sophisticated model isn’t automatically the best model for a particular problem,” Anuyah said. The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Sydney Anuyah, Ogenna Ehiemere and Oluwarotimi Ogundele tested different modelling approaches using Bitcoin and Ethereum market data as part of their master’s capstone project. The researchers compared statistical methods, including Vector Autoregression (VAR), with Long Short-Term Memory (LSTM), a deep-learning model used for analysing sequential data. According to Anuyah, the LSTM recorded strong results in some of the experiments but its performance varied, while the statistical approaches produced more consistent results and were easier to interpret. “The most sophisticated model isn’t automatically the best model for a particular problem,” Anuyah said. The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

The researchers compared statistical methods, including Vector Autoregression (VAR), with Long Short-Term Memory (LSTM), a deep-learning model used for analysing sequential data. According to Anuyah, the LSTM recorded strong results in some of the experiments but its performance varied, while the statistical approaches produced more consistent results and were easier to interpret. “The most sophisticated model isn’t automatically the best model for a particular problem,” Anuyah said. The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

According to Anuyah, the LSTM recorded strong results in some of the experiments but its performance varied, while the statistical approaches produced more consistent results and were easier to interpret. “The most sophisticated model isn’t automatically the best model for a particular problem,” Anuyah said. The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“The most sophisticated model isn’t automatically the best model for a particular problem,” Anuyah said. The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

The research used market data at 15-minute intervals and sought to produce directional trading signals rather than rely only on forecasts of exact future cryptocurrency prices. The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

The approach allowed the researchers to classify market conditions into possible buy, sell or hold decisions as new data became available. Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Anuyah said the researchers considered model performance alongside factors such as stability, interpretability, computational cost and the amount and quality of data available. “I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“I would rather have a model that performs consistently and whose behaviour I understand than a model that occasionally produces exceptional results but is much less stable,” he said. The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

The finding is limited to the models, data and conditions examined in the capstone and does not establish that statistical models generally perform better than deep-learning models in cryptocurrency trading. The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

The team also backtested its trading strategy using historical Bitcoin data. In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

In one simulation covering January 1 to September 30, 2023, the strategy generated a return of about 123 percent, according to Anuyah, compared with a 62.2 percent increase in Bitcoin over the same period. Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Anuyah added, “This is historical backtesting. A 123 percent return in an experiment does not mean someone should expect a 123 percent return in the future.” Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Backtesting allows researchers to apply a strategy to historical market data to examine how it would have performed. Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Its results can differ from live trading, where transaction costs, market conditions and other factors can affect returns. For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

For the researchers, the test also examined whether identifying the likely direction of a market could provide useful information without requiring an accurate prediction of its exact future price. “Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“Most users don’t necessarily need another complicated forecast sitting on a screen. “They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“They need information that helps them make a decision,” Anuyah stated. Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Anuyah’s work on the project followed several years working with data and financial technology. A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

A University of Lagos graduate, he worked at Edan Investments before joining Chaka Technologies in 2021 on a two-month data annotation contract. He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

He later moved into data engineering at the company and subsequently worked as a Growth Specialist. He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

He enrolled in WorldQuant University’s Financial Engineering programme while working with financial and market data at Chaka. He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

He explained, “At Chaka, my exposure to financial data became much deeper. “I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“I was spending a lot of time working with and thinking about financial markets, looking at movements, patterns and trying to understand what the numbers were actually telling us.” Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

Anuyah later relocated to the United States for graduate studies in Applied Data Science, where his work has expanded into artificial intelligence and natural language processing. He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

He said he is interested in examining how information outside conventional price and trading data, including financial news and market commentary, could be incorporated into quantitative models. “You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share

“You can build a model with excellent predictive accuracy and still have a poor trading strategy,” Anuyah said. Related News Lagos: Food innovation no longer optional to feed megacity Group writes Soludo, seeks end to indiscriminate taxes, levies in Anambra FCT Council chairs again shun Reps’ summons over N100bn audit query Share