Posts
Trading Range
1 hour ago
Thursday, July 2, 2009
OT: Kitesurfing
As a brief break from talking about trading, I recently had my bachelor party at Jetty Island in Everett where my brother and another friend snapped some shots of me kitesurfing.
Monday, June 22, 2009
Fooled by Randomness
Just picked up the book Fooled by Randomness by Nassim Taleb. I'll post a review when I've finished it.
Monday, June 8, 2009
Remembering the Hits and Forgetting the Misses
In my last post, I talked about how people underestimate the probability of things happening by coincidence. I also talked about how this causes people to read more meaning into coincidental events than what is really there.
In that post I briefly alluded to another common phenomena in human psychology...how people tend to remember hits and forget misses. This creates psychological bias and leads people to believe in connections and other things that don't really exist.
Have you ever thought of a particular person in your life, and a minute later the person calls you on the phone? Often we tend to think such a coincidence is eerie and maybe that there's some sort of supernatural psychological connection.
But how many times did you think of that person and the person didn't call? Probably too many times to count. But you forget all these times, and the one time they did call, it's such a powerful coincidence that you remember it. You remember the "hit", and forget the "misses", leading you to believe in a connection that really isn't there.
This phenomena is what leads people to believe in things like astrology, psychics, or "speakers to the dead." Have you ever watched John Edward's "The Other Side"? Here's a guy who claims to be able to "hear" the dead relatives of others, and even will show his "powers" on national TV. But what John is really doing is a technique called "Cold Reading." He is basically using high probability guesses to learn more information about the dead relative. He uses cues of the live person he is speaking to, to get an idea if he is right or wrong. He emphasizes the guesses he gets right and quickly moves on from the guesses he gets wrong.
An uncritical person watching the show will notice John's hits, and forget John's misses. This leads the person to believe that John truly has the ability to speak to the dead.
If you ever watch the show, count how much John actually gets wrong. I've done this before and he gets quite a bit wrong. It's important to pay attention to the misses and not just the hits.
How does this relate to trading?
I wrote some past blogs about designing your own trading system. I talked significantly about backtesting. When you backtest your trading system, it's important to be aware of the phenomena of "remembering hits/forgetting misses."
Let's say you've got a certain type of trade setup, and you've made some huge profits a few times off this type of setup. You start thinking that maybe you've got a nice setup that will work well.
However, how many times have you seen the same setup and you took a loss? Do you remember or notice those times?
In my own trading, I had some certain types of setups that I thought were good trading setups. However, over time I noticed I wasn't as consistently profitable as I thought I should be with them. What was the problem?
I paid more attention to the times where I profited off those setups, versus the times where I didn't. Also, when I backtested, it was easier to notice the setups that would've resulted in a profit versus the ones that didn't. My tendency to remember the hits created a skewed idea of how good the trading setups were.
So, it's important to be aware of ALL of the times you notice a particular trade setup, and how often you truly profit off those setups.
In that post I briefly alluded to another common phenomena in human psychology...how people tend to remember hits and forget misses. This creates psychological bias and leads people to believe in connections and other things that don't really exist.
Have you ever thought of a particular person in your life, and a minute later the person calls you on the phone? Often we tend to think such a coincidence is eerie and maybe that there's some sort of supernatural psychological connection.
But how many times did you think of that person and the person didn't call? Probably too many times to count. But you forget all these times, and the one time they did call, it's such a powerful coincidence that you remember it. You remember the "hit", and forget the "misses", leading you to believe in a connection that really isn't there.
This phenomena is what leads people to believe in things like astrology, psychics, or "speakers to the dead." Have you ever watched John Edward's "The Other Side"? Here's a guy who claims to be able to "hear" the dead relatives of others, and even will show his "powers" on national TV. But what John is really doing is a technique called "Cold Reading." He is basically using high probability guesses to learn more information about the dead relative. He uses cues of the live person he is speaking to, to get an idea if he is right or wrong. He emphasizes the guesses he gets right and quickly moves on from the guesses he gets wrong.
An uncritical person watching the show will notice John's hits, and forget John's misses. This leads the person to believe that John truly has the ability to speak to the dead.
If you ever watch the show, count how much John actually gets wrong. I've done this before and he gets quite a bit wrong. It's important to pay attention to the misses and not just the hits.
How does this relate to trading?
I wrote some past blogs about designing your own trading system. I talked significantly about backtesting. When you backtest your trading system, it's important to be aware of the phenomena of "remembering hits/forgetting misses."
Let's say you've got a certain type of trade setup, and you've made some huge profits a few times off this type of setup. You start thinking that maybe you've got a nice setup that will work well.
However, how many times have you seen the same setup and you took a loss? Do you remember or notice those times?
In my own trading, I had some certain types of setups that I thought were good trading setups. However, over time I noticed I wasn't as consistently profitable as I thought I should be with them. What was the problem?
I paid more attention to the times where I profited off those setups, versus the times where I didn't. Also, when I backtested, it was easier to notice the setups that would've resulted in a profit versus the ones that didn't. My tendency to remember the hits created a skewed idea of how good the trading setups were.
So, it's important to be aware of ALL of the times you notice a particular trade setup, and how often you truly profit off those setups.
Saturday, June 6, 2009
The Probability of Coincidence
If you go over to Tim Sykes's site, you'll see a blog over there about Hemispherx Biopharma (HEB). The attention this blog has attracted is quite amusing. In summary, a bunch of people are accusing Tim of being cohoots with Adam F. of TheStreet.com. Why do they think this? Well, Adam came out with an article slamming HEB at 8 AM EST. Tim sent out a TIMAlert to his subscribers at 10:20 AM EST that he had shorted HEB. Adam's article was reposted around this time, making some people think that Adam and Tim were involved in some type of well-timed conspiracy to bring down HEB and to cause the huge tank that was seen on Friday.
Of course, this is all just coincidence, but it shows people's tendencies to read way more into things than what are usually there. In fact, it is a well documented that humans underestimate the probability of things happening by coincidence. It is this human tendency that causes people to believe in conspiracy theories such as the fake moon landing, the JFK assasination, or the idea that 9/11 was an inside job.
Coincidental events, even strange or incredible ones, are actually quite common. This is easily illustrated by the birthday paradox. The birthday paradox shows how, if you have a group of randomly selected people, the probability that two of those people will have the same birthday is quite high. In a room of 23 people, the probablity is over 50%. In a room of 57 people, the probability is over 99%. Thus, if you're in a room with a lot of people and you meet someone with the same birthday as you, there's no inner or hidden meaning behind it....it's just basic probability at work.
Given that there are 5 billion people on this planet, and that there are billions of various events that occur, and that there are 5 million minutes in a 10 year period, it is bound that coincidences, even amazing ones, will happen quite often. In fact, the most incredible coincidence would be if there were no coincidences.
When it comes to the Tim HEB situation, it's obvious that everything that happened was just a coincidence. This is definitely a case where Occam's Razor applies. The simplest explanation, consistent with all the facts, is the most likely explanation. What is more probable here? That Tim & Adam are part of some well-timed conspiracy to bring down HEB? Or that HEB had a big run-up and just attracted a lot of attention from traders, two of which happened to write about it? Obviously, given the thousands of traders out there, and the fact that some of them are writing/blogging, the probability of a coincidental event is quite high....particularly with a stock like HEB that had run up so far so fast and has attracted a lot of attention.
If you believe the logic of the conspiracy theorists, then that would mean that anyone who shorted HEB on Friday and happened to mention it on the internet to other traders could've been part of the conspiracy. I shorted HEB multiple times on Friday. If I had tweeted that I had shorted HEB, would this mean that I was part of a conspiracy too? Often you can see the absurdity of conspiracy theories just by critically examing them. In fact, most conspiracy theories show their absurdity by examining the sheer number of people that would often have to be involved....logic dictates that large conspiracies are nearly impossible given that the more people that are involved, the more likely the whistle will be blown and the conspiracy will fail before it ever gets off the ground.
This relates to my past blog where I wrote how humans often see patterns where there are none. My next blog post will be about another human tendency....the tendency to remember "hits" and forget "misses"....I'll talk about how this creates psychological bias and how it relates to trading.
Of course, this is all just coincidence, but it shows people's tendencies to read way more into things than what are usually there. In fact, it is a well documented that humans underestimate the probability of things happening by coincidence. It is this human tendency that causes people to believe in conspiracy theories such as the fake moon landing, the JFK assasination, or the idea that 9/11 was an inside job.
Coincidental events, even strange or incredible ones, are actually quite common. This is easily illustrated by the birthday paradox. The birthday paradox shows how, if you have a group of randomly selected people, the probability that two of those people will have the same birthday is quite high. In a room of 23 people, the probablity is over 50%. In a room of 57 people, the probability is over 99%. Thus, if you're in a room with a lot of people and you meet someone with the same birthday as you, there's no inner or hidden meaning behind it....it's just basic probability at work.
Given that there are 5 billion people on this planet, and that there are billions of various events that occur, and that there are 5 million minutes in a 10 year period, it is bound that coincidences, even amazing ones, will happen quite often. In fact, the most incredible coincidence would be if there were no coincidences.
When it comes to the Tim HEB situation, it's obvious that everything that happened was just a coincidence. This is definitely a case where Occam's Razor applies. The simplest explanation, consistent with all the facts, is the most likely explanation. What is more probable here? That Tim & Adam are part of some well-timed conspiracy to bring down HEB? Or that HEB had a big run-up and just attracted a lot of attention from traders, two of which happened to write about it? Obviously, given the thousands of traders out there, and the fact that some of them are writing/blogging, the probability of a coincidental event is quite high....particularly with a stock like HEB that had run up so far so fast and has attracted a lot of attention.
If you believe the logic of the conspiracy theorists, then that would mean that anyone who shorted HEB on Friday and happened to mention it on the internet to other traders could've been part of the conspiracy. I shorted HEB multiple times on Friday. If I had tweeted that I had shorted HEB, would this mean that I was part of a conspiracy too? Often you can see the absurdity of conspiracy theories just by critically examing them. In fact, most conspiracy theories show their absurdity by examining the sheer number of people that would often have to be involved....logic dictates that large conspiracies are nearly impossible given that the more people that are involved, the more likely the whistle will be blown and the conspiracy will fail before it ever gets off the ground.
This relates to my past blog where I wrote how humans often see patterns where there are none. My next blog post will be about another human tendency....the tendency to remember "hits" and forget "misses"....I'll talk about how this creates psychological bias and how it relates to trading.
Tuesday, May 19, 2009
Saturday, May 9, 2009
I just applied at DNDN
I just applied for a part-time medical writer position at Dendreon Corporation in Seattle. Right after I applied I realized that it was DNDN....a volatile stock that has made for some great trades in the past.
Funny how, in the trading world, you often only know companies by their ticker symbols....
Funny how, in the trading world, you often only know companies by their ticker symbols....
Friday, May 8, 2009
Rough Week But Profitable
This past week was a rough week but I ended the week with a small profit.
I started off the week strong. On Monday I had my best trading day thus far with over a $400 profit. However, I was obviously aided by the strong market day. The next day was a wash with a $15 loss.
Then came Wednesday. It seemed like nothing went right. I had very few winning trades. I got caught in failed breakouts. I made mental errors. I chased some trades. I went long on some preleaders near the open despite my post last week that I don't do as well when I do that. When I did something wrong, I didn't get away with it, and when I did something right, it wouldn't go my way. And then I had flat out bad luck. For example, I went long on DAN at $1.88. DAN crossed $2 a couple times. I should've taken a profit when it failed to hold above $2 the second time, but I wasn't watching it when that happened. Suddenly DAN tanked all the way to $1.80 and my profit turned into a loss.
I also missed opportunities by trying too hard to save myself on commissions. Since I do direct access trading with Cy Group, I have ECN fees that can add to my trading costs. If I use limit orders that aren't immediately fillable, I can get a rebate which can cut my commissions in half. The only problem with this is that, if you're momentum trading like I do, sometimes you just need to get in and not worry about the ECN fee. I had 2-3 trading opportunities (don't remember the tickers) that day that I missed because, rather than hitting the offer (or bid for a short) to get in right away, I sat on the bid (or sat on the ask for a short) to try to get the ECN rebate. Of course my order never got filled and the stocks zoomed away in my intended direction.
All in all, Wednesday was over a $400 loss which wiped out my excellent Monday.
Thursday was a small losing day. Again I made some of the same mental errors that I made Wednesday....trying to get save my commissions (and missing a few trade opportunities), and again going long on a few preleaders near the open.
Today I was determined not to make the same errors. I stopped trying to save myself commissions and just hopped into trades if the opportunity was there. I also avoided longing the preleaders near the open. I still made a few mental errors today but nothing like Wednesday or Thursday. Today I had over a $200 profit to give myself about a $125 profit for the week. Not a great week compared to my previous weeks but at least it's a profit.
Here's my P/L by day for the week. This graph was generated by TradePerformance.
Here's the stats. These stats are nothing like the previous weeks.
Report Generated by TradePerformance™ Software - http://www.TradePerformance.com
Copyright (c) 2000-2008 SCG Associates, LLC. All Rights Reserved
TradePerformance is a Trademark or Registered Trademark of SCG Associates, LLC in the United States and other countries
One thing that stands out about these stats is that I did better on my shorts than my longs. In fact, I was a loser on my longs, which hadn't been the case in any of the previous weeks. This is not a function of market conditions because 2 of the days this week were strong market days. In fact, today was a strong market day yet I was more profitable on my shorts today than my longs.
Being the analytical person I am, I was curious if there were certain times of the day where I did better with my longs compared to other times of the day. I was mainly interested in just the morning because it seems like the morning makes or breaks my trading day. On my most profitable trading days, I had a good start to the morning, while on my worst days, I usually started off in the hole and had to try to climb back out.
I went back and looked at my trading stats for going long on the day rangers. I broke my stats up by 15 minute time frames. Here's my overall stats for the past 4 weeks for trading these:
6:30 - 6:45 AM PST: 17 winners 13 losers 57% winning percentage
6:45 - 7:00 AM PST: 9 winners 14 losers 39% winning percentage
7:00 - 7:15 AM PST: 10 winners 6 losers 63% winning percentage
7:15 - 7:30 AM PST: 10 winners 4 losers 71% winning percentage
There definitely appears to be a time of day effect. I do better than a coin flip near the open, but then in the second 15 minutes of the trading day I don't do very well. Then in the second half hour the performance returns (and is in fact better than near the open).
My hunch is that, when I go long the second 15 minutes, I'm more likely to get caught in a stock that's already made a move and is ready to pull back.
Another thing I've noticed is that I'm not letting my long profits run enough. Normally I like to lock in half when I'm 10% up on my position (or $1 up for a stock trading over $10). However, I'm finding the day rangers often go much further than this. CATY was one example today. I went long at $14.06. I set an order to sell half at $14.97 (near the round $ mark....a bit under in case it didn't quite make it there). It hit that, then went over $15. It hit around $15.20ish then started to pull back. I got out more at $15.26 and $15.17 when it started pulling back. However, it then went on to $16. I'm considering a different profit taking system then what I have. Rather than the partial profit method, I will simply set a target and wait for the stock to pass that target. Once it's passed the target, I will take profits on a 10/60 cross to the downside on the 1-minute chart. I've done significant back-testing with this and it would result in much greater profits that what I've been getting. With CATY I would've been out at $15.90 using this system.
So here's the things to work on next week:
1. Don't go long on day rangers the second 15 minutes of the trading day
2. Don't get cute with trying to get ECN rebates. I've lost a lot more money in missed trades than the small amount of money I would've saved in rebates. If the opportunity is there, just get in.
3. Work on letting winners run more with a slightly different profit-taking system.
Because of my struggles this week, I will not be increasing my position sizes next week. I need more consistency and discipline first.
I started off the week strong. On Monday I had my best trading day thus far with over a $400 profit. However, I was obviously aided by the strong market day. The next day was a wash with a $15 loss.
Then came Wednesday. It seemed like nothing went right. I had very few winning trades. I got caught in failed breakouts. I made mental errors. I chased some trades. I went long on some preleaders near the open despite my post last week that I don't do as well when I do that. When I did something wrong, I didn't get away with it, and when I did something right, it wouldn't go my way. And then I had flat out bad luck. For example, I went long on DAN at $1.88. DAN crossed $2 a couple times. I should've taken a profit when it failed to hold above $2 the second time, but I wasn't watching it when that happened. Suddenly DAN tanked all the way to $1.80 and my profit turned into a loss.
I also missed opportunities by trying too hard to save myself on commissions. Since I do direct access trading with Cy Group, I have ECN fees that can add to my trading costs. If I use limit orders that aren't immediately fillable, I can get a rebate which can cut my commissions in half. The only problem with this is that, if you're momentum trading like I do, sometimes you just need to get in and not worry about the ECN fee. I had 2-3 trading opportunities (don't remember the tickers) that day that I missed because, rather than hitting the offer (or bid for a short) to get in right away, I sat on the bid (or sat on the ask for a short) to try to get the ECN rebate. Of course my order never got filled and the stocks zoomed away in my intended direction.
All in all, Wednesday was over a $400 loss which wiped out my excellent Monday.
Thursday was a small losing day. Again I made some of the same mental errors that I made Wednesday....trying to get save my commissions (and missing a few trade opportunities), and again going long on a few preleaders near the open.
Today I was determined not to make the same errors. I stopped trying to save myself commissions and just hopped into trades if the opportunity was there. I also avoided longing the preleaders near the open. I still made a few mental errors today but nothing like Wednesday or Thursday. Today I had over a $200 profit to give myself about a $125 profit for the week. Not a great week compared to my previous weeks but at least it's a profit.
Here's my P/L by day for the week. This graph was generated by TradePerformance.
Here's the stats. These stats are nothing like the previous weeks.
Trading Statistics - Selected Trades
| | All Trades | Long Trades | Short Trades |
| Net Profit | 124.64 | (108.71) | 233.35 |
| Total Gains | 2061.94 | 1525.76 | 536.18 |
| Total Losses | (1937.295) | (1634.465) | (302.83) |
| Ratio of Gains to Losses | 1.064 | 0.933 | 1.77 |
| Number of Trades | 140 | 103 | 37 |
| Number of Winning Trades | 74 | 53 | 21 |
| Number of Losing Trades | 66 | 50 | 16 |
| Percentage of Winning Trades | 52.857% | 51.456% | 56.757% |
| Percentage of Losing Trades | 47.143% | 48.544% | 43.243% |
| Largest Winning Trade | 75.76 | 75.76 | 53.04 |
| Largest Losing Trade | (112.00) | (112.00) | (51.17) |
| Average Winning Trade | 27.86 | 28.79 | 25.53 |
| Average Losing Trade | (29.353) | (32.689) | (18.93) |
| Average Trade | 0.890 | (1.055) | 6.31 |
| Win/Loss Ratio | 0.949 | 0.881 | 1.35 |
| Average Winning Trade Length | 0.00 | 0.00 | 0.00 |
| Average Losing Trade Length | 0.00 | 0.00 | 0.00 |
| Maximum Consecutive Winners | 9 | 6 | 6 |
| Maximum Consecutive Losers | 8 | 7 | 7 |
| Expectancy | 0.8903 | (1.0554) | 6.3068 |
| Expectancy Percent | 3.0330% | (3.2287)% | 33.3217% |
| Mathematical Expectation | 0.030 | (0.032) | 0.333 |
| Kelly Maximum Bet Size | 0.086 | (0.005) | 0.323 |
| Kelly Maximum Reduced 50% | 0.0428 | (0.0027) | 0.1617 |
Copyright (c) 2000-2008 SCG Associates, LLC. All Rights Reserved
TradePerformance is a Trademark or Registered Trademark of SCG Associates, LLC in the United States and other countries
One thing that stands out about these stats is that I did better on my shorts than my longs. In fact, I was a loser on my longs, which hadn't been the case in any of the previous weeks. This is not a function of market conditions because 2 of the days this week were strong market days. In fact, today was a strong market day yet I was more profitable on my shorts today than my longs.
Being the analytical person I am, I was curious if there were certain times of the day where I did better with my longs compared to other times of the day. I was mainly interested in just the morning because it seems like the morning makes or breaks my trading day. On my most profitable trading days, I had a good start to the morning, while on my worst days, I usually started off in the hole and had to try to climb back out.
I went back and looked at my trading stats for going long on the day rangers. I broke my stats up by 15 minute time frames. Here's my overall stats for the past 4 weeks for trading these:
6:30 - 6:45 AM PST: 17 winners 13 losers 57% winning percentage
6:45 - 7:00 AM PST: 9 winners 14 losers 39% winning percentage
7:00 - 7:15 AM PST: 10 winners 6 losers 63% winning percentage
7:15 - 7:30 AM PST: 10 winners 4 losers 71% winning percentage
There definitely appears to be a time of day effect. I do better than a coin flip near the open, but then in the second 15 minutes of the trading day I don't do very well. Then in the second half hour the performance returns (and is in fact better than near the open).
My hunch is that, when I go long the second 15 minutes, I'm more likely to get caught in a stock that's already made a move and is ready to pull back.
Another thing I've noticed is that I'm not letting my long profits run enough. Normally I like to lock in half when I'm 10% up on my position (or $1 up for a stock trading over $10). However, I'm finding the day rangers often go much further than this. CATY was one example today. I went long at $14.06. I set an order to sell half at $14.97 (near the round $ mark....a bit under in case it didn't quite make it there). It hit that, then went over $15. It hit around $15.20ish then started to pull back. I got out more at $15.26 and $15.17 when it started pulling back. However, it then went on to $16. I'm considering a different profit taking system then what I have. Rather than the partial profit method, I will simply set a target and wait for the stock to pass that target. Once it's passed the target, I will take profits on a 10/60 cross to the downside on the 1-minute chart. I've done significant back-testing with this and it would result in much greater profits that what I've been getting. With CATY I would've been out at $15.90 using this system.
So here's the things to work on next week:
1. Don't go long on day rangers the second 15 minutes of the trading day
2. Don't get cute with trying to get ECN rebates. I've lost a lot more money in missed trades than the small amount of money I would've saved in rebates. If the opportunity is there, just get in.
3. Work on letting winners run more with a slightly different profit-taking system.
Because of my struggles this week, I will not be increasing my position sizes next week. I need more consistency and discipline first.
Subscribe to:
Posts (Atom)