ML4T FINAL PRACTICE QUESTIONS ANSWERED CORRECTLY LATEST UPDATE 2026

ML4T FINAL PRACTICE QUESTIONS ANSWERED CORRECTLY LATEST UPDATE 2026 Why did it become a good investment to bet against mortgage-backed securities. A) The default rate on the mortgages kept rising. B) Rating agencies were accurately assigning ratings. C) Banks were incentivized to issue more and more mortgages. D) A and C - Answers D) A and C In terms of the fundamental law, when aiming for high performance, you can make up for low skill with ___. A) low breadth B) low sharpe ratio C) high breadth D) high volatility - Answers C) high breadth Is Bollinger Bands a leading indicator? A) Yes B) No, Bollinger Bands is a lagging indicator C) No, Bollinger Bands is not an indicator D) No, Bollinger bands is neither a leading or a lagging indicator - Answers B) No, Bollinger Bands is a lagging indicator What should you do when the stock hits the bottom Bollinger Band? A) Hold your position (don't change your position at all) B) Exit any position you currently have and then hold at 0 shares C) Short the stock to have negative shares D) Long the stock to have positive shares - Answers D) Long the stock to have positive shares Who was the first bank Michael approached to make him the CDSs? A) JP Morgan B) Goldman Sachs C) Deutsche Bank D) Bank of America - Answers B) Goldman Sachs If the Information Coefficient is doubled and the trading opportunities are multiplied by 9, what happens to the Information Ratio. A) The Information Ratio is multiplied by 1 (remains unchanged) B) The Information Ratio is multiplied by 4.5 C) The Information Ratio is multiplied by 6 D) The Information Ratio is multiplied by 18 - Answers C) The Information Ratio is multiplied by 6 What is the best way an investor who is lacking in skill can beat a more skillful investor? A) Hire skillful people B) Make smarter trades C) Make more trades per day D) Pick better stocks - Answers C) Make more trades per day Which statement is true regarding the Fundamental Law of Portfolio Management? A) IR = IC * sqrt(BR) B) IC = IR * sqrt(BR) C) BR = IC * IR D) IR = IC + sqrt(BR) - Answers A) IR = IC * sqrt(BR) Which of the following can not improve your performance based on the Fundamental Law? A) Improve your skill B) Increase the Breadth C) Increase the IC and BR D) Decrease breadth but improve the skill - Answers D) Decrease breadth but improve the skill What does CDS stand for? A) Collateralized Debt Swap B) Credit Default Swap C) Collateralized Debt Security D) Credit Default Security - Answers B) Credit Default Swap In the movie "The Big Short", what did Credit Default Swaps do? A) Insured the bonds against failure and the insurer paid the claim if the underlying bonds fail B) Insured the bonds against failure and the insurer paid the claim if the underlying bonds make money C) To increase employees D) As an annual practice - Answers A) Insured the bonds against failure and the insurer paid the claim if the underlying bonds fail Which of the following is not the keys of Grinold's fundamental law? A) Skill B) Performance C) Luck D) Breadth - Answers C) Luck Which of these scenarios does not indicate an upward trend in stock price: A) A particular stock's price surpasses the upper Bollinger Band B) A particular stock's 20-day SMA surpasses its 100-day SMA C) A particular stock's short-term EMA falls below its long-term EMA D) The value of a particular stock's Bollinger Band Percentage 0.8 and its price/SMA ratio is = 1 - Answers C) A particular stock's short-term EMA falls below its long-term EMA Which of these could represent a State in a Reinforcement Learning problem when referring to a trading scenario? A) BUY B) Daily Return C) Return from Trade D) All of the above - Answers B) Daily Return Which celebrity chef and "Kitchen Confidential" author explains collateralized debt obligations (CDOs) by comparing them to fish stew? A) Gordon Ramsay B) Mario Batali C) Thomas Keller D) Anthony Bourdain - Answers D) Anthony Bourdain Following statement is true or false: Hedge Funds cannot succeed if EMH applies. A) True B) False - Answers B) False In Reinforcement Learning, the default MDP has an assumption of infinite horizons to overcome that, we introduce a concept of ___________ rewards. Multiplying the reward by λ raised to t. Where λ's limits are ___ λ =___. A) [reduced , -1, 1] B) [reduced, 0 , 1] C) [discounted, -1, 1] D) [discounted, 0 , 1] - Answers D) [discounted, 0 , 1] Do you think that a market that is semi-strong efficient is also weak form efficient? Why or Why not? A) Yes it does, because semi-strong EMH states that the prices adjust immediately to all publicly available information, which also includes the asset's price history and vol data. B) No it does not, because semi-strong EMH talks only about fundamental data such as the company's earnings. C) Yes it does, because semi-strong EMH states that the prices adjust immediately to all publicly and privately available information. D) No it does not, as semi-strong EMH talks about information that is mutually exclusive to weak form EMH. - Answers A) Yes it does, because semi-strong EMH states that the prices adjust immediately to all publicly available information, which also includes the asset's price history and vol data. Given that EMH is about 3 forms, weak, semi-strong, and strong. What would be the behavior of prices, if, for example, the standard investor became one that only bought and held stocks passively? A) There would be no change in the way that prices adjusted B) Prices will fail eventually to show any type of new information. There will be incentives to trade more C) Prices will continue to change based on new information that won't cease to enter the market D) We would probably enter another recession - Answers B) Prices will fail eventually to show any type of new information. There will be incentives to trade more The goals of unsupervised learning and reinforcement learning are: A) The goal of unsupervised learning is to find similarities and differences between data points; the goal of reinforcement learning is to find a suitable action model to maximize cumulative reward; B) The goal of unsupervised learning is to find a suitable action model to maximize cumulative reward; the goal of reinforcement learning is to find similarities and differences between data points; C) The goal of both unsupervised learning and reinforcement learning is to find similarities and differences between data - Answers A) The goal of unsupervised learning is to find similarities and differences between data points; the goal of reinforcement learning is to find a suitable action model to maximize cumulative reward; What is the meaning of "synthetic CDO" in "the big short" movie? A) A CDO that contains subprime mortgages B) A CDO that contains student loan C) A CDO that contains credit default swaps D) A CDO that contains corporate debt - Answers C) A CDO that contains credit default swaps According to the Fundamental Law, which of the following statements is most accurate? A) Increasing skill and/or breadth will increase performance, but it's more difficult to increase breadth than skill, and increasing skill isn't as effective as increasing breadth. B) Increasing skill and/or breadth will increase performance, but it's more difficult to increase skill than breadth, and increasing breadth isn't as effective as increasing skill. C) Utilizing machine learning will increase performance. D) A Georgia Tech education including CS7646 coursework will increase performance. - Answers B) Increasing skill and/or breadth will increase performance, but it's more difficult to increase skill than breadth, and increasing breadth isn't as effective as increasing skill. Suppose you purchase a call option for a single stock (assume that you can do so) for $10 on a day where the stock price is $7 under the strike price. Starting from then, the price rises monotonically through the expiration date, on which the price of the stock is $3 above the strike price. What is the maximum net profit that you can make during this time period, assuming your only choices are to either exercise or not exercise the option? A) -$10 B) -$7 C) -$3 D) $3 - Answers B) -$7 Which of the following statements about Dyna is true? A) Dyna is a model-free based method. B) Dyna is the basis for Q-Learning. C) Dyna is a blend of model-free and model-based methods D) Dyna is more expensive than Q-Learning - Answers C) Dyna is a blend of model-free and model-based methods