Varying Lot Sizes

Did not manage to achieve much with machine learning, so I’m putting a pause on that first.

Spent some time analysing the effects of varying lot sizes.

Here are six strategies I tested:

1. Use a fixed lot size of 1 unit
2. Reduce lot size by 0.1 unit if the past three trades are losers (subject to a minimum lot size of 0.25 unit); increase lot size by 0.1 unit if the past 1 trade is a winner (subject to a maximum lot size of 1 unit)
3. Reduce lot size by 0.1 unit if the sum profit of the past three trades are losers (subject to a minimum lot size of 0.25 unit); increase lot size by 0.1 unit if the past 1 trade is a winner (subject to a maximum lot size of 1 unit)
4. Reduce lot size by 0.1 unit if the past two trades are losers AND both trades exit within 10 days or less (subject to a minimum lot size of 0.25 unit); increase lot size by 0.1 unit if the past 1 trade is a winner (subject to a maximum lot size of 1 unit)
5. Filter using a sector ETF (XLK for NVDA and XLE for SWN). For buy trades, the ETF’s previous week close must be above the 50 Week EMA. For sell trades, the ETF’s previous week close must be below the 50 Week EMA.
6. Filter using items 4 and 5 above. When considering the past two trades for item 4, even trades that are filtered out by item 5 are considered.
7. Filter using items 4 and 5 above. When considering the past two trades for item 4, trades that are filtered out by item 5 are NOT considered.

Here are the results:

Stock		1	2	3	4	5	6	7
NVDA	Final	$535,775.20	$519,712.67	$319,602.16	$444,114.22	$603,843.98	$495,345.77	$534,484.08
	Lowest	$98,552.63	$98,552.63	$98,552.63	$98,552.63	$99,504.50	$99,504.50	$99,504.50
	Max DD	-21.71%	-20.70%	-15.61%	-19.12%	-10.10%	-7.65%	-9.05%
SWN	Final	$102,577.56	$109,679.01	$110,343.65	$114,660.12	$129,416.21	$124,188.28	$134,847.54
	Lowest	$60,035.70	$67,214.40	$78,946.28	$72,273.37	$73,467.27	$81,467.93	$79,460.22
	Max DD	-39.96%	-32.79%	-21.05%	-27.73%	-26.53%	-18.53%	-20.54%

The results above are based on a 5% risk per trade, with a starting capital of $100,000.

Analysis

Two stocks is too small a sample size to say anything definitively. However, the good thing about these two stocks is that one of them represents a trending stock (NVDA) and the other represents a non-trending stock (SWN).

Overall, in terms of reducing the maximum drawdown, strategy 6 works the best, but this strategy requires us to keep track of trades that are not taken.

For instance, if I spot two potential sell trades but the previous week’s ETF close for that sector is above the 50 Week EMA, I should not take the trade. However, I should monitor the trades to track if they will end up being profitable. If they are both not profitable and exit within 10 days, I should reduce the lot size for the next trade. In other words, lot size for the next trade is calculated ‘as if’ I took those two trades.

This can be confusing and difficult to track. In addition, it can cause psychological stress if I see that the trades I missed end up being very profitable. Monitoring it closely can lead to FOMO for subsequent trades. Based on my previous practice sessions, I realize that missing a good trade can mess me up quite a bit. Hence, strategy 6 might not be the most suitable strategy for me.

Taking the above into consideration, strategy 7 seems to be a good option. It is easier to implement and also leads to higher profits for both SWN and NVDA. The only drawback is that the maximum drawdown is slightly higher than strategy 6.

Side Note – Days to Exit

On a side note, based on my practice results for NVDA and SWN, the maximum number of days it takes for me to exit a trade is 35, the average is about 10 days. This is great news because that means I can likely still use options to trade. I may have to roll out in some cases, but in most cases, I can exit within slightly more than one trading month.