# Tag Info

15

Because of: The (extreme) dominance of noise over signal The prevalence of non-repeating patterns (many of which we know are not going to repeat) A pathetic sample size for cross-validation Regime changes due to exogenous events. These are typically in the cross-val window which makes it even worse. (GFC, financial integration, trade law changes, interest ...

7

I recommend reading Cao, Hansch, and Wang (2004) "The Informational Content of an Open Limit Order Book". They present a simple model for an order-book price called the weighted price ($\mbox{WP}$): $$\mbox{WP}^{n_1 - n_2} = \frac{\sum_{j=n_1}^{n_2} (Q_j^d P_j^d + Q_j^s P_j^s)}{(Q_j^d + Q_j^s)}$$ Where: $n$ is the order book level $Q_j$ is the size at ...

6

You can pull a list of all stocks easily. See this question. You can get nasdaqlisted.txt and otherlisted.txt from here. nasdaqlisted.txt is clearly Tape C. otherlisted.txt contains an Exchange column which can be used to determine Tape A or B. If it is N it's listed at NYSE and therefore Tape A, otherwise it's Tape B. Also, NYSE publishes a symbol list ...

5

Few points from my experience: 1 Another filters that you that you should consider is for price = 999 or 999.99 that appears in some data providers. 2 Another set of checks is to look at cross-section of e.g. range = (high-low)/close over all names. Check for the smallest range and largest range to see if the values make sense. You can also check daily % ...

3

I was going to comment but it turned out to be quite elaborate. My experience with certain AI/ML methods is that they're not deterministic. Take RBM for instance, a very wide-spread paradigm. To train such a machine you have two approaches, backpropagation or Kullback-Leibler divergence. Both require you to initialise the machine to a random state. And ...

3

Definitely time series analysis. What you essentially want to do is some form of impact analysis. this can be done naturally using multivariate time series models like Vector Auto Regression models. Also when working with data to model liquidity you might want to use some specialized procedures like GARCH and ACD. Further there are methods to model non ...

3

Your first definition is wrong; I'm not sure where you got that from. Your second definition is correct: the ISO alerts the exchange that the submitting party has taken responsibility for RegNMS and requests a fill at only that venue's price; there is no routing away. Obviously, there is a huge red-tape burden to get permission to do this.

3

I can only repeat myself because your mentioned previously asked question is essentially identical: => I would say do not include non-trading days, do not include days with zero position, do not include days where the asset did not trade for whatever other reason. Here some reasons and pointers: Sharpe measures excess returns scaled by volatility. The ...

3

Tape A is NYSE-listed stocks. Tape C is NASDAQ-listed stocks. Tape B is the regionals, ie. everything else. Most of Tape B is on ARCA now days, though some of it is still on the NYSE MKT, formerly known as AMEX. If you want to use Google Finance, just note that they prepend every symbol with the exchange its on. For example, if I search for "AAPL", Google ...

2

The basic CAPM - which is what your regression estimates - says $$R_S = R_f + \beta_S (R_{Market}-R_f)$$ where $$\beta_S = \frac{Cov(R_M,R_S)}{Var(R_M)}$$ i.e. the return of a certain stock depends only on the correlation with the market portfolio. For your pricing equation to work, you will need to have an idea about the expected market (excess) ...

2

Margin is determined by those who are in the business of settling options contracts. In the OTC that would be those who make markets in OTC options. (Subject to regulatory requirements if imposed) IV is determined by market forces, supply and demand. The same applies to options written on bananas and chimps. liquidity depends again on supply and demand of ...

2

Having developed many custom backtesting programs in the past, I wish I would have just started out by purchasing a decent commercial backtester for a few hundred dollars. The cost of buying one already completed will save you hours of time on learning the nuances and problems that come with backtesting implementation. Once you are adept at the ins and outs ...

2

The adjusted close will change after dividends and stock splits. So the old data will have to be replaced by the new. So it is usually a good idea to check for adj close of the downloaded values against current values. I also like to check for downloaded data against some other source (like Google). I do this by writing a unit test that will randomly pick a ...

2

As someone pointed out, most times you do not have a choice because how the order book is disseminated heavily depends on the exchange. In Asia there are exchanges that even show the name of the counter party (though only aggregate volume not each order). Other exchanges provide a full "view", again others price level aggregates. Broadly speaking, some ...

2

The main problem with hypothesis testing is going to be survivorship bias. Any manager with a track record you're looking at is only there because they haven't performed badly -- if they perform badly then investors withdraw their money, they collapse, and you don't have their data to do the hypothesis test on! So even if all the managers were investing in ...

2

It's simply the difference between the two values. High-Low gives you an idea about the total price movement over the tick period, Open-Close gives you an indication of the direction of the move. If Open-Close is negative, the price went up, and vice versa. High-Low obviously is always positive.

2

Breakpoint approaches Test based To be well received in a financial econometrics journal, you want test-based approaches. Depending on your question it is common to see a linear regression (least squares) where the parameter suspected of breaking is interacted with an indicator function $I(E)$ where $E$ is the event in question; this function assumes a ...

2

I am not aware about any rule of thumb on market impact during fixing auctions (nothing comparable with the usual square root rule for meta orders market impact). In Navigating Liquidity 6 - A global menu for optimal trading, there is a study about the liquidity during the fixing auctions, you should start from here. Another important point is the data you ...

2

You are not doing anything wrong. You just need to multiply the absolute return by the currency conversion factor. Example: You trade 200,000,000 yen notional and generate a return of 16% on that notional, then simply multiply 32,000,000 jpy gain by your conversion factor 0.0126 to yield a return of 403,200 USD. The return of 16% was generated on the ...

2

It's probably because of the strong long-standing statistical underpinnings in economics and econometrics, and overall, risk prediction. For example, look at current research with fat-tail distributions and calculations for Expected Tail Loss (ETL), etc. These studies fit Student's t, Normal, Stable, and Pareto probability distributions to data and report ...

1

This answer can certainly be improved with more information: like which instruments, what time scale etc. If you can assume one instrument to be the 'base' instrument then the ratio of the prices is a good measure with both time series beginning at the same time. This is similar to calculating relative return. I have used this when backtesting a pairs ...

1

That's a great question and it is what I always wanted to try to do. I guess I found a solution using PDE approach. Change of numeraire would be more intuitive indeed, but I am not very good in stochastic calculus. The idea is as follows: 1) Let's consider portfolio $\Pi = V(X,Y,t) - \Delta_X X - \Delta_Y Y$. I will found $\Delta_X$ and $\Delta_Y$ such ...

1

No you certainly should not use in simulation any data that you don't know in real environment because simply results will be much different. Simulation should be created in environment as much possible similar to real environment. You basically should use only Open values for making trades, High/Close/Low you can use for indicator calculation of past ...

1

Its a bit of a broad and almost philosophical question to be honest. Just two things that pop in mind are (A) the money supply (& possible inflation) in combination with interest rates - this will drive the major money allocation in the world. As you see now with S&P at record highs on the back of massive quantitative easing and low interest rates. ...

1

I think Hull treats dilution in his book, and it's extensible to this case. For what it's worth, the strike is typically set low enough that there's little doubt about exercise, meaning there's not much point in modeling the optionality. Most people concentrate on the dilution alone -- particularly the question of the change in fully diluted versus ...

1

Beside the lag issue, I do not believe that the negative correlation between Consumer Staples and sentiment is counter-intuitive. The S&P SPDR XLP ETF (Consumer Staples) offers a broad exposure to defensive mega-cap consumer names. Spending on consumer staples is usually characterized by an inelastic demand. Whether the economic climate is strong or ...

1

I'm currently also using daily returns which I want to annualize. This is my approach: For every month, I calculate the simple return using the formula: (end-of-month closing price / beginning-of-month closing price) - 1. I use the Excel formula somproduct(geomean(A1:A12+1)-1) to find the monthly compounded return. Finally, I annualize the result of step 2 ...

1

The key reason why you observe divergent performance patterns is related mostly to the following: The biggest reason is the different cost to hedge those products. The costs to implement and especially maintain the hedge on the long vs short side can be very different. Either the hedge is implemented through an index replication in which case the manager ...

1

There are three prices to consider when discussing an ETF: the ideal price as represented by the index, the NAV of the fund based on that day's holdings, and the market value traded on a stock exchange. This third price is what you see. In your example, Russell has calculated a cap-weighted value based on the annual membership. Direxion then determines ...

1

I recommend to check tradestation.com Their EasyLanguage allows to build and backtest much more advanced strategies. They also have pre-build indicators and data for many years. It is not cheap, about \$100 a month but most of the time they have a special offers for first 3 or 6 month free and monthly fee is dropped if you reach a certain level of trading. ...

Only top voted, non community-wiki answers of a minimum length are eligible