# Final Flashcards Preview

## CS 7646 - OMSCS > Final > Flashcards

Flashcards in Final Deck (64)
Loading flashcards...
1
Q

What is a ‘tick’ in stock aggregation?

A

A successful buy/sell match. The smallest resolution in stock data aggregation.

2
Q

Are prices at different exchanges guaranteed to match each other?

A

No

3
Q

How is stock data aggregated? (time frames, columns)

A

Typically minute by minute or hour by hour chunks. Each chunk has the following entries:

• open - first transaction price
• high - highest prices during time period
• low - lowest prices during time period
• close - last transaction price
• volume - total volume over time period

In this class we will always use time chunks of 1 day.

4
Q

Describe stock splits

A

Your single share splits into multiple shares.

Examples: a 4:1 split means you get 4 shares for every 1 of yours.

The value of the company doesn’t actually change though, because the price of each stock also decreases by the same ratio.

5
Q

Why do stocks split?

A

Because the price is too high.

Reasons high prices are bad:

• Trading in options (batches of 100 shares for instance) becomes very difficult
• Difficult to buy even one share if the price is too expensive
• Hard to fine tune a portfolio if stock prices are too high
6
Q

How do you calculate the adjusted close?

A

You work backwards in time, dividing the stock value by the stock split ratio from when it occurred, backwards until the beginning.

So if stock split like:

• 2015-05-01 - 2:1
• 2013-05-01 - 4:1

At the very end of time, adjusted close == close (that will always be true) and will be true moving backwards until you hit the first split. @ 2015-05-01 you divide the close by 2. Do that for all time backwards until you hit the next split at 2013-05-1. Then divide it by 2 and by 4 (so by 8). Do that until the next split, etc.

7
Q

How can dividends affect stock prices?

A

When the dividend is announced, the price will start to climb up to an increased value + dividend value. When the dividend is paid, the price will go back down to it’s price minus the dividend value.

8
Q

How do you account for dividends in the adjusted close price?

A

You subtract off the value of the dividend from the day it was paid for all time backwards.

9
Q

What is survivor bias?

A

When looking at historical data and selecting stocks for your portfolio, you can’t only use stocks that exist at the end of your data range. This is because those stocks are the one that ‘weathered the storm’ - they made it to the end whereas others may have not.

If you use stocks that existed at the beginning if your data, you will likely find decreased overall performance. That difference between using original and final stocks is the bias. You can purchase survivor bias free data to account for this.

10
Q

Will the adjusted close always be the same regardless of what years data you are looking at?

A

No - if you have a file from 2012 and 2015, it’s likely there were stock splits/dividends paid since 2012. So if you looked at the adjusted close of a date in 2010 in the 2012 file, it would likely be higher than the same 2010 date found in the 2015 file (because in the 2015 file you’re accounting for additional splits and dividends).

11
Q

What are the assumptions made by the EMH (efficient markets hypothesis)?

A
• large numbers of investors
• New information arrives randomly
• Prices adjust quickly
• Prices reflect all available information
12
Q

Where does information come from?

A
• Price/Volume: public, basis for technical analysis
• Fundamental: public, comes out quarterly
• Exogenous: Info about the world that affects the company (eg, price of oil compared to airlines)
• Company Insiders: secretive, reflects info you have that ppl outside company don’t have
13
Q

What are the forms of the EMH?

A
• weak: future prices can’t be predicted by analyzing historical prices and volumes (current price reflects all info that is known, so tech analysis doesn’t help, eg history doesn’t affect the future). Leaves room for Fundamental Analysis
• semi-strong: prices adjust rapidly to new public information. Prohibits Fundamental Analysis b/c when quarterly reports come out, prices react immediately
• strong: prices reflect all public and private information. Essentially impossible to make money by holding a portfolio other than an index.
14
Q

Is the EMH Correct?

A

Price To Earnings Ratio (P/E, low is better), tends to refute the semi-strong EMH because over many decades of analysis, P/E was a strong predictor of earnings, which is a result of Fundamental Analysis.

15
Q

What is arbitrage? (Ch 8)

A

Buying a security in one market and simultaneously selling it in another at a higher price. Price difference could be due to exchange rates having not been incorporated into price or difference information being available in each market.

16
Q

What does the word “efficient” mean in the context of Efficient Market Hypothesis, EMH? (Ch 8)

A

How quickly relevant information moves through the marketplace. The more efficient a market, the more that relevant information is equally available to all participants.

17
Q

In what markets is the EMH likely strongest? (Ch 8)

A

Large markets that are highly transparent and liquid (Like large cap stocks in the US). Studies have shown in these markets very few investors consistently outperform the market index.

18
Q

What is a reverse split? (Ch 12)

A

Literally the opposite of a split. A stock priced at \$30 before a 1 for 3 split would then be valued at \$90.

19
Q

What is the Fundamental Law of Active Portfolio Management, conceptually (equation)? (Ch 9)

A

performance = skill * sqrt(breadth)

• skill: how well managers turn information about an equity into accurate prediction of future returns
• breadth: number of trades manager makes each year
20
Q

What are two ways to increase the breadth of a portfolio? (Ch 9)

A
1. Hold more equities at once

2. Turn your current equities over more frequently

21
Q

What are the two broad categories of portfolio management risk? (Ch 9)

A
1. Systematic Risk: A risk undertaken by exposure to any asset in the asset class (eg, if the asset class fails, so d you b/c you’re part of that asset class)
2. Specific Risk: Risk associated with a specific asset. “Oil company prices suffers if new oil field fails to produce as expected”
22
Q

What type of risk can diversification help against? (Ch 9)

A

Specific risk

Std of portfolio returns declines as more individual stocks are included in the portfolio.

23
Q

What are the issues with diversification? (Ch 9)

A
• It has diminishing returns - > 20-40 different stocks no longer helps
• You can know less depth about each stock, the more stocks you’re managing (eg, alpha takes effort to learn)
24
Q

What is risk-adjusted return? (Ch 9)

A

Returns / std

25
Q

How is the risk-adjusted reward like the sharp ratio? (Ch 9)

A

It’s the same thing except it doesn’t account for the risk free return. If you set risk free return to 0, then the sharp ratio is equal to the risk-adjusted return.

26
Q

What is the Information Ratio? (Ch 9)

A

Risk adjusted reward. Calculated as:

IR = IC * sqrt(BR)

IR = expected reward / std(reward)

IR(alpha) = mean(alpha_p[t]) / std(alpha_p[t])

IR = (port return - benchmark return)/ std(port return - benchmark return)

This is focused on “excess” return, which is why beta is not included.

Also excess because it’s the excess over the benchmark returns. It’s trying to measure skill and take out the market effects.

r_p(t) = beta_p * r_m(t) + alpha_p(t)

27
Q

What types of risk are presented by the CAPM model? (Ch 9) - [types of risk]

A
• Market risk: std(beta * market return)

- Investor specific, or “skill” risk: std(alpha)

28
Q

What is the Information Coefficient? (Ch 9)

A

Correlation of a managers predictions about asset prices with the actual future prices.

```1 = perfectly correct
-1 = perfectly wrong```
29
Q

What is the REAL Fundamental Law of Active Portfolio Management (equation)? (Ch 9)

A

Information Ratio = Info. Coefficient * sqrt(breadth)

30
Q

What makes up a Markov Decision Process (MDP)? (03-05)

A
• States (S)
• Actions (A)
• Transition Function (T[s, a, s’])
• Reward Function (R[s, a])
31
Q

Do you need the Transition and Reward functions for Q Learning?

A

No

32
Q

What is an experience tuple?

A
• when in state s
• and you take action a
• you end up in state s’
• and received reward r
33
Q

What is a finite vs infinite horizon in RL?

A

Effectively it’s what gamma represents.

Technically your horizon is 1 / (1-gamma), but an an infinite horizon is when gamma = 1 and a finite horizon is when gamma < 1.

34
Q

How is trading mapped into RL?

A
• Market = Environment
• Actions = Buy, Sell, hold, etc
• States = Features (indicators)
35
Q

What is a stock option?

A

A contract which gives buyer a legal right but not the obligation to buy/sell the underlying stock at a specific price (strike price) on or before the expiration date.

(In europe you can’t exercise the option before the expiration date, only on it)

36
Q

What in an option strike price?

A

The price you are guaranteed to be able to buy at regardless of the price the root (stock) symbol is trading at.

37
Q

How many shares does an option control?

A

It’s always 100! But the stock prices is prices per share.

38
Q

Why would you use an option over a stock?

A
• You don’t have to tie up the full amount it would cost you to buy 100 shares of the stock (leverage!)
• Your risk of loss is limited to no more than the option cost
39
Q

What are the downsides of options trading?

A
• The premium is lost money, period.
• Expiration dates - once it’s expired, it’s expired. You can’t wait for for the stock to go up at a later date.
• You don’t own the stock
• person who sold you the option still owns the stock. They collect dividends, have voting rights, etc.
40
Q

What’s the intrinsic value of a stock (with respect to options)?

A

For a call:

Stock price - strike price

For a put:

Strike price - stock price

41
Q

What does ‘in-the-money’ mean (options)?

A

If you were to buy an option and exercise it immediately, then you would make money (this basically is the intrinsic value).

42
Q

What does `out-of-the-money` mean (options)?

A

If you bought an option and exercise it immediately, you would lose money (no intrinsic value)

43
Q

What is the time value (options)?

A

The excess premium cost of an option beyond its intrinsic value.

Generally the more time that remains until expiration , the higher the time value

44
Q

What is time decay (options)?

A

The rate at which an option is currently losing its time value. (theta) Options prices approach their intrinsic value as they approach expiration.

45
Q

What is a put option?

A

The opposite of a call. You’re buying the right to sell the stock at the strike price.

You short the stock. Borrow 100 shares, sell those at the strike price. Buy 100 shares at market rate (lower than strike), and give back to borrower so you net the difference. The max profit is limited b/c the lowest possible prices is 0.

46
Q

What is a covered call and what are the three ways in which it can work?

A

Buy a stock and write a call on that same stock. The ‘safest’ thing you can do.

stock ends above strike:
- stock is called away; sell at strike, lose profit above
profit: Strike - (purchase price) + premium
now we don’t have the stock anymore

stock ends up, but below strike:
- perfect; option not exercised
profit: current price - purchase + premium
made money on stock and you still have it

stock goes down:
- option not exercised, but still have stock
profit: current price - purchase + premium
premium partially offsets the loss

47
Q

What composes the premium of an option?

A

Premium = intrinsic value + time value

48
Q

Are bonds lower risk than a blend of stocks and bonds?

A

No - Harry Markowitz showed a blend of stocks and bonds is actually lower risk than bonds only.

49
Q

What kind of covariance should two stocks have in order to minimize risk?

A

Closer to -1 is better. When one goes up, the other goes down and vice versa. Assuming they have the same CR in the end, they would have the lowest overall portfolio volatility.

50
Q

What kind of correlation do you want between two stocks in your portfolio over the short and long term?

A

Negative in the short term and positive in the long term.`

51
Q

What is mean variance optimization (MVO)?

A

An optimization technique that considers variance and covariance and determines an optimal set of portfolio weights to hit a target return while minimizing risk.

52
Q

What are the thresholds for target return in MVO?

A

It must be between the min and max returns of the all of the possible assets in the portfolio.

53
Q

What are the inputs and outputs to MVO?

A

Inputs:

• Expected Return - how you think each stock will perform in the future
• volatility - std dev of historical DR of each stock
• covariance - covariance matrix between each stock and every other stock’s DR
• target return - the return you’d like to have.

Outputs:
- Set of weights that should try to meet your target return while minimizing overall volatility

54
Q

How would the risk of an MVO optimized portfolio compare to risk of each individual stock in the portfolio?

A

It would be lower than the lowest risk.

55
Q

What is the efficient frontier?

A

If you were to plot a line of MVO optimized portfolios for varying target returns you get some kind of line.

You can’t have a portfolio above that line, and any portfolio below that line is generally leaving something on the table, such as getting lower return for the same risk or taking on more risk for the same return.

Anything on the line is considered most efficient.

It’s not actually used that much in practice because it’s rather theoretical, but it’s often used as a gauge to see where your efficiency may be.

56
Q

How do you calculate the max possible sharpe ratio of an efficient frontier curve?

A

Draw a line from the origin to be tangent with the efficient frontier curve. That is the point with the maximum sharpe ratio.

57
Q

What are the loss and profit limits of a buy call?

A

max loss: premium
max profit: inf

• losing premium at price <= strike
• Break even at strike + premium
• profit at price > strike + premium
58
Q

What are the loss and profit limits of a buy put?

A

max loss: premium
max profit: strike price - premium (price can’t go < 0)

• losing premium at price >= strike
• Break even at strike - premium
• profit at price < strike - premium
59
Q

What is a write call/put?

A

call: Selling someone the option to buy my stock at strike price until expiration.
put: Selling someone the option to sell you stock at the strike price until expiration.

60
Q

What are the loss and profits of a write call?

A

max loss: inf
max profit: premium

• profit premium at price <= strike
• Break even at strike + premium
• losing money at price > strike + premium
61
Q

What are the loss and profits of a write put?

A

max loss: strike - premium (price can’t go below 0)
max profit: premium

• profit premium at price >= strike
• Break even at strike - premium
• losing money at price < strike - premium
62
Q

What percent of stocks are not exercised?

A

90%! This is why writing options even with large loss potential may not be so risky.

63
Q

How can you profit from a sideways market?

A

Using the Butterfly (Options)!

• Buy two calls, one higher and one lower than current price.
• Write two calls half way between the upper and lower bounds of your two calls.

You make max profit when the price is at the value of your two calls.

It tails off in either direction until a break even and then it flatlines at the max loss of your entry price.

64
Q

What is tammer’s advice to new investors?

A
1. Look for theoretically sound ideas, things based on economics and finance
2. It should be empirically tested
3. It should be simple, otherwise it’s likely overfitting