Judgmental Biases
In the second chapter of Trade Your Way to Financial Freedom, Tharp talks about the mental filters that we use to parse the vast amounts of information with which we are bombarded everyday. These judgemental heuristics allow us to reach conclusions which often coincide with our already established notions, making it difficult to re-examine our conclusions. Tharp breaks these biases down into those that affect system development, testing systems, actually trading a developed system.
System Development
Representation bias is when "people assume that when something is supposed to represent something, it really is what it is supposed to represent." Tharp goes on to give an example of a daily bar chart. Of course the bar represents the day's movement, but its really a line on paper, not the market itself. You would need to know volume, the number of new/old positions were opened/closed, the number of trades that took place and at what prices, the number/identity of large institutional or small retail investors placed trades, the number of long/short positions, opinions of market movement, etc. to truly know what the market was doing that day.
In short, it would take a lot of information and time to get an accurate picture of the market on a given day. So naturally we used a line on a chart(s) to try to distill all of that information into a managable item. Even if you could process all of the information in its raw form, what would you do with it?
Reliability bias is assuming that the data we use is reliable. I know for myself that it often is not. I remember looking at FX prices from Oanda side-by-side with those for the same currency and time period from MetaTrader. They did not agree. I've also seen beta calculations and P/E ratio calculations for stocks from various free online sources and seen a disparity. Vendors use different methods to produce their results and so what they produce may not be as reliable as you might like.
Lotto bias is having a feeling of increased confidence from manipulating data, feeling that your maniuplation will lead you to control over the market's movements. I'm extremely guilty of this one. I often think that if I can find the right series of manipulations to any given set of data, I can make a really good guess at what will happen. Its a sneaky psychological effect that says "because I have seen some information and put my special touch on it, I can control what will happen." I guess the name of this bias comes from playing a lottery where you pick the numbers, feeling as if picking the "right" numbers increases your odds of winning.
The bias of the law of small numbers is when you see a few examples of a particular methodology working and generalize that this particular methodology is successful. Unfortunately you don't spend the time to examine it in all circumstances to see how truly successful it is. I'm guilty of this based on looking at charts for technical cues. I love to try to read candles, and when I spot a few successful patterns in the books or even in real charts, I think they absolutely work.
The conservatism bias is when, after establishing an idea in our mind, we are unwilling to examine conflicting evidence or ideas. I'm sure that I'm guilty of that one, or at least I don't have a visceral reaction telling me that I never exhibit this bias.
The randomness bias is a belief that market prices move in a random fashion and the erroneous assumptions about what that randomness might mean. Tharp argues that while academic studies have shown that markets are random (I'm thinking of Malkiel's Random Walk Down Wall Street - I need to read that one day), there are extreme tails in the market's distribution that would not arise from normally distributed random prices. Tharp further argues that the greater extremes in price movement lead traders to underestimate risk. Tharp goes on to say that if markets are random, then traders misinterpret what that randomness means. When they see a long trend in a price series, their natural inclination is to create a theory to explain the trend, which would lead to seeking patterns and establishing causal relationships. This also leads to trying to pick tops and bottoms in the market.
This one probably pertains to me in that I have been obsessed with finding the highs and lows in markets, trying to find the turning points. In fact, the last FX system I tried developing was to find candlestick patterns for turning points. Needless to say, it didn't perform well.
The need to understand bias is definitely a sore spot for me. This relates to the need to theorize on why markets are doing what they are doing. I've had countless conversations expressing confusion at various market behaviors, usually in the face of some bit of fundamental news, like a better than expected earnings report. I had a conversation just this week about the stock market's sell off in the face of news that the Fed will continue to leave interest rates at a record practically zero level. I think the financial media has an awful lot to do with this, as everyday they tell us "why" the market moved the way it did, and they do it as if they have analyzed the situation and can pinpoint exactly the reason. So we think there is always a reason, and we always go looking for it. Of course by "we" I mean "I".
I'll get into the other biases in the next post.
System Development
Representation bias is when "people assume that when something is supposed to represent something, it really is what it is supposed to represent." Tharp goes on to give an example of a daily bar chart. Of course the bar represents the day's movement, but its really a line on paper, not the market itself. You would need to know volume, the number of new/old positions were opened/closed, the number of trades that took place and at what prices, the number/identity of large institutional or small retail investors placed trades, the number of long/short positions, opinions of market movement, etc. to truly know what the market was doing that day.
In short, it would take a lot of information and time to get an accurate picture of the market on a given day. So naturally we used a line on a chart(s) to try to distill all of that information into a managable item. Even if you could process all of the information in its raw form, what would you do with it?
Reliability bias is assuming that the data we use is reliable. I know for myself that it often is not. I remember looking at FX prices from Oanda side-by-side with those for the same currency and time period from MetaTrader. They did not agree. I've also seen beta calculations and P/E ratio calculations for stocks from various free online sources and seen a disparity. Vendors use different methods to produce their results and so what they produce may not be as reliable as you might like.
Lotto bias is having a feeling of increased confidence from manipulating data, feeling that your maniuplation will lead you to control over the market's movements. I'm extremely guilty of this one. I often think that if I can find the right series of manipulations to any given set of data, I can make a really good guess at what will happen. Its a sneaky psychological effect that says "because I have seen some information and put my special touch on it, I can control what will happen." I guess the name of this bias comes from playing a lottery where you pick the numbers, feeling as if picking the "right" numbers increases your odds of winning.
The bias of the law of small numbers is when you see a few examples of a particular methodology working and generalize that this particular methodology is successful. Unfortunately you don't spend the time to examine it in all circumstances to see how truly successful it is. I'm guilty of this based on looking at charts for technical cues. I love to try to read candles, and when I spot a few successful patterns in the books or even in real charts, I think they absolutely work.
The conservatism bias is when, after establishing an idea in our mind, we are unwilling to examine conflicting evidence or ideas. I'm sure that I'm guilty of that one, or at least I don't have a visceral reaction telling me that I never exhibit this bias.
The randomness bias is a belief that market prices move in a random fashion and the erroneous assumptions about what that randomness might mean. Tharp argues that while academic studies have shown that markets are random (I'm thinking of Malkiel's Random Walk Down Wall Street - I need to read that one day), there are extreme tails in the market's distribution that would not arise from normally distributed random prices. Tharp further argues that the greater extremes in price movement lead traders to underestimate risk. Tharp goes on to say that if markets are random, then traders misinterpret what that randomness means. When they see a long trend in a price series, their natural inclination is to create a theory to explain the trend, which would lead to seeking patterns and establishing causal relationships. This also leads to trying to pick tops and bottoms in the market.
This one probably pertains to me in that I have been obsessed with finding the highs and lows in markets, trying to find the turning points. In fact, the last FX system I tried developing was to find candlestick patterns for turning points. Needless to say, it didn't perform well.
The need to understand bias is definitely a sore spot for me. This relates to the need to theorize on why markets are doing what they are doing. I've had countless conversations expressing confusion at various market behaviors, usually in the face of some bit of fundamental news, like a better than expected earnings report. I had a conversation just this week about the stock market's sell off in the face of news that the Fed will continue to leave interest rates at a record practically zero level. I think the financial media has an awful lot to do with this, as everyday they tell us "why" the market moved the way it did, and they do it as if they have analyzed the situation and can pinpoint exactly the reason. So we think there is always a reason, and we always go looking for it. Of course by "we" I mean "I".
I'll get into the other biases in the next post.
posted by Young Pip at 1:36 PM
0 Comments:
Post a Comment
<< Home