Message From TraderJanie
This is Part II of a discussion about the new “Tick Size Pilot Program (TSPP)” if you’d like to review Part I click here. Today I am going to finish off that discussion. This is something you should all be aware of and not ignore, although I have to admit it’s all kinda boring. To be truthful it not kinda boring, it is boring. It is not the kind of reading you’d do for entertainment but, if you consider yourself a trader, it’s information YOU NEED TO KNOW. So please bear with me and let’s finish this off together. I promise a more interesting submission the following week.
Pessimism or Optimism
Most people have heard the question – “Is the glass half empty, or half full ?” What we want to do in this section is to first explain a simple approach to answering some of the questions we posed in the first part. Then we want to model a ”worst case,” or pessimistic scenario in order to assess the impact that the Trade Size Pilot Program may have in the future.
Rounding is important to understand
However, we need to digress for a minute and talk about rounding. We are all probably familiar with standard, arithmetic rounding. For example, assuming an increment of 0.01, the number 12.014 rounds to 12.01, and the number 12.015 rounds to 12.02.
There are other types of rounding, though. A number can be rounded up to the nearest increment, or rounded down to the nearest 0.01 increment. For example, rounding up either 12.014 or 12.015 up to the nearest 0.01 yields 12.02. Rounding either number down to the nearest 0.01 increment yields 12.01.
Before answering those questions of Part 1, though, let’s also outline a simple model. Let’s say that we have two “actors” in this scenario
The market and market makers are going to set the bid and ask prices. Let’s assume that they have full access to the order book and know what price that crosses would occur at but are allowed to price certain stocks in 5 cent increments. A pessimistic (but realistic) view is that we can model the market/market makers’ actions as being pessimistic for the trader. Why? Because it is in the market/market makers best interest to do the following:
For example, if a market maker could normally do a cross of a buy order at 12.03, he might round up to 12.05 for a stock with a 5 cent increment that he sells, because this incents him to provide more liquidity at that price.
As we noted in the previous newsletter, we are primarily concerned with answering two questions:
1. How will we get the price target (limit) with the correct increment for trades to enter?
2. How can we model various systems and how they would have performed in the past under this program so that we can project the impact of TSPP into the future?
With that in mind, let’s dispatch question #1 quickly. Here at Quantitrader we will look up trades that are posted each night in the current FINRA group list, and pessimistically round the targets for those tickers that will trade in 5 cent increments – based on the order type of the limit order.
But, be aware that your broker must follow the current FINRA group list, also
Most brokers will enforce the price increment in their order screens. They will undoubtedly make every effort to conform to this program, but we should probably not expect 100% accuracy – especially with a changing group constituency.
Question #2 from above is a very important question. Recall that we backtest models in order to gain confidence that the trading system will provide the desired performance in various market environments. With the Trade Size Pilot Program, we do not know the makeup of the group constituency prior to Oct. 3, 2016. So backtesting presents a unique problem. We have to address it, though, in order to know that we have a good system going forward under the TSPP. If we take the most pessimistic approach in modeling and get an acceptable degradation in performance, then we can have confidence that our basic model will thrive.
So, in addition to rounding all target orders as outlined in the table above, we are going to run a backtest and artificially adjust the open, high, low, and close prices that the market might have offered for all of the tickers that are currently in group G1, G2, or G3 using 5 cent increments.
We are going to do this in a pessimistic manner from the standpoint of the trader. For example, when he buys, the quote prices will be rounded up. And, when he sells, the quote prices will be rounded down. Note again that we cannot know the past constituency of the 3 groups, but using the current constituency extended over the past year should yield a useful approximation.
Here is the bottom line for Ultra backtesting for 6 months with 1 cent rounding for all trade targets versus a backtest with 5 cent “pessimistic” rounding for price targets and quotes for all tickers that are currently in the 3 groups.
|1 cent rounding||5 cent rounding|
8% of trades were affected.
So, regarding question #2, we can conclude from the table above, that this would have had a minimal negative effect on performance. The system will perform well going forward with TSPP
Check our out the Performance Matrix for our algorithms and sign up today.