Trading Stats - What do you measure?

Quote from Grob109:


If you are a beginner approximate it at 11.1 % every 6.6 days on your chart starting with your present capital. Shift it upward by 20 % if you spend a half day working the screens with me using your method.
More...

I am a beginner in my trading equities. The problem for me is to have enough flawless candidates(what i consider flawless is of course subjective) to enter through out the month.

My current stock universe is around 200 stocks. With my current universe, i have been able to trade around 5 stocks per month yielding around 10% per month.

This month has been better given I have been in BOOM, REDF, and RHAT with 2 very good cycles and 1 ok cycle.

You mention shifting performance up 20% if someone were able to connect with you. I was hoping you would elaborate on how one would go through the process of doing this.

I did have one question regarding the % increase. Are you shifting the cycle which is 6.6 days to 20% more or are you shifting the monthly percent by 20%?

My current solution for increasing profits per month was to get a universe in the range of 500 to monitor rather than 200 which would hopefully allow me to have many more flawless stocks to enter. I am sure their are much better ways to improve performance.

Thanks for the great post.
 
There are, of course, ways of measuring the smoothness of the equity curve. If you're running thousands of computerized what-if scenarios and don't want to visually inspect each one of the equity curves, you can use one or more of these "how smooth is it" measurements to sieve out the most promising dozen or two, then look only at those.

If you're mathematically inclined you can invent your own (oh my god, that means you have to THINK!) ways of crunching the data of an equity curve and deriving a smoothness measure.

Other people have tackled this problem and have invented quite a number of smoothness-of-equity-curve measurements, which you may or may not like:
  • Kestner's K-Ratio
  • The Ulcer Index
  • The Lake Ratio
  • R-squared (fitting a straight line to the equity curve)
  • The Semideviation Ratio
All of which you can look up on the web. Several of them are available as built-in, or user add-on, modules in Wealth-Lab and Trading Blox Builder.
 
Quote from Grob109:


......
I guess what I do is out in right field as usual.

I have been keeping track of the quality of stocks in my universe. This assures that they are price performers and also "repeatable". That is they do the same thing over and over.

I continually monitor the place in the trading cycle they occupy as time passes. This way I know three things continually: where they are; what is next; and how fast the cycle is operating in time.

I rank my stocks by their contemporary money velocity (%gain in price per day. Think of it at two piles of stocks where what I own is on the left with the top rank on the bottom of the pile. On the right is the hot list that I have from me universe with the best money velocity on the top. I continually compare what is on the top of each pile.

I cross trade from the owned to the hot list with money I make available buy selling the owned and buying the hot.

The money velocity of a stock in a cycle peaks as the price goes through the axis of the cycle. This means I am buying stocks that are increasing in money velocity as they approach the axis and selling stocks as they decrease in money velocity after they cross the axis. Typical lower values of money velocity in this market are 4.5 %/day. A three day hold would be like going from 19 to 28 for a better money velocity stock. See the last two runs of IRIS. In a similar range see the last five runs of BOOM.
......
More...

Is right field the correct place to be ? ....

Most of your posts are not decipherable: this one is.

This is exactly what need s to be done to keep trading profitably. Perhaps it could be stated in fewer words like trade volatility and reproducible patterns .....

I dont trade the same issues if they fail my tests and I change the mix of markets and issues according to observable conditions.

Once an issue breaks down you need to move on until things improve.
 
A couple of other measurements I look at in both backtesting and actual results:

1) average drawdown , which is the daily average of (max equity minus current equity) over the entire period -- gives a feel for the frequency and length of drawdowns

2) expectancy minus standard error -- helps remind me to be suspicious of results with too few trades and/or too much inconsistency
 
Quote from tradingbug:

My current solution for increasing profits per month was to get a universe in the range of 500 to monitor rather than 200 which would hopefully allow me to have many more flawless stocks to enter.
More...

Have you thought of starting with a daily computer scan of the entire stock market? You could still focus on just 200 (or fewer), but they could be different ones each day.
 
my question is very elementry but here goes...From the books that i have read on trading, they talk about using excel spreadsheets, but i do not understand exactly why. i read about trade stats. but why about the actual price data? Is there a method of understanding the relationship between the prices, ex. O, H, L , C? If there is what is it called ? Excuse me for my ignorance:D
 
You must log in or register to reply here.
Back
Top