Below, I will analyze the results of the trading data.
TRADING DATA
First, here are the links to the trading data. (Note that I didn't trade every day between May and July 2009. When I didn't feel that there were any good opportunities that day, I simply took the day off).
Broker Statements, May 2009 to July 2009
Trading Data Posts
May 4, 2009
May 6, 2009
May 7, 2009
May 8, 2009
May 11, 2009
May 12, 2009
May 13, 2009
May 14, 2009
May 15, 2009
May 18, 2009
May 19, 2009
May 20, 2009
May 21, 2009
May 22, 2009
May 26, 2009
May 27, 2009
May 28, 2009
May 29, 2009
June 1, 2009
June 2, 2009
June 3, 2009
June 4, 2009
June 5, 2009
June 8, 2009
June 9, 2009
June 10, 2009
June 12, 2009
June 15, 2009
June 16, 2009
June 17, 2009
June 18, 2009
June 19, 2009
June 22, 2009
June 23, 2009
June 24, 2009
June 25, 2009
June 26 to 30, 2009
July 1 to 14, 2009
July 15 to 16, 2009
July 17, 2009
July 21, 2009
July 23, 2009
July 24, 2009
July 27, 2009
SUMMARY OF TRADING DATA, MAY TO JULY 2009
Non-compounded data:
Gain Win/Loss
May 35.39% 29 - 7
June 67.44% 29 - 10
July 36.39% 6 - 0
139.22% 64 - 17
220 - 53 when w/l sample is adjusted to reflect the actual number of trades made
ANALYSIS OF TRADING DATA, MAY TO JULY 2009
May Trades
The May trades range from May 4 to May 29. Those are the settlement dates. The actual trade dates were Apr 29 to May 26.
The trades listed do not list each filled order separately. When creating my transaction history, my broker groups several orders together. That is what appears on my posts.
My personal trading records show I actually made 215 trades during that period.
Although I made 215 trades in May (which comes to 108 buy/sell combinations), the broker's records posted show only 36 buy/sell combinations (29-7).
If you translate my 80.56 win/loss percentage from the 36 combinations to the 108 combinations, you get 87/108. So, my actual win/loss record in May was likely to be about 87/108 (80.56%).
The percentage gains posted show a gain of 35.39% during May (which may be understated due to broker errors, as explained at the end of this post).
But 35.39% significantly understates the percentage gain I actually earned, because I was compounding the gains. Every time I made a gain (or loss), I traded with the new capital level. Therefore, my actual dollar gain during May would've been significantly higher than 35.39%. My actual broker documents show the compounded dollar gain, but I've hidden the dollar amount for privacy reasons and because I want to demonstrate the percentage gains one can earn using my methods when non-compounded.
How much did compounding increase my earnings? I'm not able to recreate it exactly, since I don't know the exact trade figures for all 108 buy/sell combinations, but I should be able to estimate it pretty closely.
Since I earned 35.39% over 108 buy/sell combinations, that's an average gain of 0.327685185% per buy/sell combination. When I did a simulation of 108 gains averaging this amount, the actual compounded gain comes to 42.38%, higher than the non-compounded 35.39%.
June Trades
The June trades range from June 1 to June 30. Those are the settlement dates. The actual trade dates were May 27 to June 25.
The trades listed do not list each filled order separately. When creating my transaction history, my broker groups several orders together. That is what appears on my posts.
My personal trading records show I actually made 250 trades during that period.
Although I made 250 trades in June (which comes to 125 buy/sell combinations), the broker's records posted show only 39 buy/sell combinations (29-10).
If you translate my 74.4 win/loss percentage from the 39 combinations to the 125 combinations, you get 93/125. So, my actual win/loss record in June was likely to be about 93/125 (74.4%).
The percentage gains posted show a gain of 67.44% during June (which may be understated due to broker errors, as explained at the end of this post).
But 67.44% significantly understates the percentage gain I actually earned, because I was compounding the gains. Every time I made a gain (or loss), I traded with the new capital level. Therefore, my actual dollar gain during June would've been significantly higher than 67.44%. My actual broker documents show the compounded dollar gain, but I've hidden the dollar amount for privacy reasons and because I want to demonstrate the percentage gains one can earn using my methods when non-compounded.
How much did compounding increase my earnings? I'm not able to recreate it exactly, since I don't know the exact trade figures for all 125 buy/sell combinations, but I should be able to estimate it pretty closely.
Since I earned 67.44% over 125 buy/sell combinations, that's an average gain of 0.539520000% per buy/sell combination. When I did a simulation of 125 gains averaging this amount, the actual compounded gain comes to 95.93%, higher than the non-compounded 67.44%.
July Trades
The July trades range from July 1 to July 24. Those are the settlement dates. The actual trade dates were June 26 to July 21.
The trades listed do not list each filled order separately. When creating my transaction history, my broker groups several orders together. That is what appears on my posts.
My personal trading records show I actually made 82 trades during that period. I will now reduce the number of trades I made from 82 to 80, to adjust for the broker errors that resulted in two artificial orders: the order overselling shares and then the buyback. I make this adjustment to be consistent with the fact that I adjusted my July 21 percentage gain from 9.21% to 27.75%, due to the same broker errors.
Although I made 80 trades in July (which comes to 40 buy/sell combinations), the broker's records posted show only 6 buy/sell combinations (6-0).
If you translate my 100.0 win/loss percentage from the 6 combinations to the 40 combinations, you get 40/40. So, my actual win/loss record in July was likely to be about 40/40 (100.0%).
The percentage gains posted show a gain of 36.39% during July (which may be understated due to broker errors, as explained at the end of this post).
But 36.39% significantly understates the percentage gain I actually earned, because I was compounding the gains. Every time I made a gain (or loss), I traded with the new capital level. Therefore, my actual dollar gain during July would've been significantly higher than 36.39%. My actual broker documents show the compounded dollar gain, but I've hidden the dollar amount for privacy reasons and because I want to demonstrate the percentage gains one can earn using my methods when non-compounded.
How much did compounding increase my earnings? I'm not able to recreate it exactly, since I don't know the exact trade figures for all 40 buy/sell combinations, but I should be able to estimate it pretty closely.
Since I earned 36.39% over 40 buy/sell combinations, that's an average gain of 0.909750000% per buy/sell combination. When I did a simulation of 40 gains averaging this amount, the actual compounded gain comes to 43.66%, higher than the non-compounded 36.39%.
COMPOUNDING ALL THREE MONTHS STRAIGHT
Above, I've compounded the results for each single month. But I haven't compounded them across all three months (which is what I actually did when I made the trades!)
The results compounded per month assume that I compounded throughout May, then started with a lower capital level in June, compounded that, then started with a lower capital level in July and compounded that.
In actuality, I kept compounding the results for every single trade from May to July!
To simulate the compound over the entire May to July period, I think the best thing to do would be to take the total non-compounded gain of 139.22% and divide it by the 273 buy/sell combinations over that period. The average gain per day comes to 0.509963370%.
If you compound that average gain over 273 combinations, the simulation shows that the actual compounded percentage gain comes to 300.95%, which is much higher than the 139.22% non-compounded gain!
If I was to compound the entire May to July period using a different method (instead of using the average gain of .509963370% for each of the 273 combinations), one might compound it using the following values: For the first 108 results use a gain of 0.327685185% (above, from May), for the next 125 results use a gain of 0.539520000% (above, from June) and for the final 40 results use a gain of 0.909750000% (above, from July).
When you run that simulation, the results are almost identical: a percentage gain of 300.75%.
SUMMARY OF HISTORICAL TRADING GAINS
The non-compounded broker reported monthly % gains are as follows:
Gain Win/Loss
May 35.39% 29 - 7
June 67.44% 29 - 10
July 36.39% 6 - 0
139.22% 64 - 17
As I mentioned, May and June figures may be understated due to the broker errors described here. (July was already partially adjusted to correct the broker errors).
The above figures don't take into account the compounding capital levels that I used.
The above win/loss sample size is understated as a result of the broker grouping multiple trades together.
When you take into account the compounding within each single month and adjust the win/loss sample size to reflect the actual number of trades made, the figures become:
Gain Win/Loss Win/Loss %
May 42.38% 87 - 21 80.56%
June 95.93% 93 - 32 74.40%
July 43.66% 40 - 0 100.00%
181.97% 220 - 53 80.59%
When you compound my percentage gains across all three months, my gain comes to 300.95% over 273 trades over only 3 months!
WHAT WAS THE EFFECT OF THE RISING MARKET ON MY TRADING RESULTS?
Here is how the DOW Jones Industrial Index performed during the period:
May (May 4 to May 29) 8212 to 8500
June (June 1 to June 30) 8500 to 8447
July (July 1 to July 24) 8447 to 9093
Although a rising market would (everything else being equal) inflate my results, it's important to note that I likely would've still earned very high gains had there been a down market over this period.
Why? Well, my market gains were astounding, they outperformed the market by many multiples.
And crucially, when the DOW fell about 0.6% in June, that was my best month of the three, with a compounded gain of 95.93% (although my win/loss % was the lowest of the three months, at 74.4%, which you might expect given that the DOW performed the worst that month).
TRADERS DIDN'T FULLY BENEFIT FROM THE RISING MARKET
Very crucially, it's important to note that about 90% of the time my trading positions are closed when the market closes at 4pm. This can be verified by looking at my broker documents. This means that I almost always had no positions that would either gap up or down when the market opened the following trading day.
What does this mean? Well, if you look at the ETF SPY (which tracks the overall S&P 500 market), you'll see that about 1/3 of the total gains from March 2009 to December 2009 resulted from a gap up, and only about 2/3 of the gains resulted from intraday movement!
This means that you wouldn't expect a trader like me to have benefited from the rising market nearly as much as a buy and hold investor, because I rarely held a position when the market tended to gap up!
Therefore, I achieved my trading gains in a much more difficult market than the DOW statistics above suggest, making my gains even more impressive!
