Len’s Journal

[LEONARDZIR]

Hi. Joined the mentor program in May. Just finished the coding section which I found difficult but very instructive. I have minimal coding experience which includes coding in Fortran some 40 years ago. Have decided to develop a mean reversion system for a variety of reasons . I plan to go long only in uptrending markets. Based on some prior trading I am defining an uptrend as a 50ma above a 200ma and both rising. Really very unsure of which mean reversion entries have an edge and would love to have some input from other members on entries that look promising. I will probably start testing a close below a standard Bollinger band and go long when price closes back in the Bollinger band with exits yet to be defined. Also will test 3 lower closes with an entry some atr from prior bars low on a limit order. I am quite excited to be part of this community.

[LEONARDZIR]

Forgot to mention I live in Sherborn, Massachusetts a small town near Boston, Massachusetts.(USA)

[JulianCohen]

I’m not too far in front of you Len. I knew nothing about mean reversion at all two months ago having always been a trend follower. I just coded things up I read about on the internet and experimented. Nick gave me some good advice that worked for me. He urged me to stick to one basic idea, keep it simple and test the hell out of it by applying different filters and exits. This didn’t give me something I could trade, but made me very aware of what the different criteria were doing, so when the time came I could easily develop a system to trade for myself.

You will find a lot of very helpful and above all smart people on this forum. Good luck and enjoy the ride.

[LEONARDZIR]

Thank you Julian

[LEONARDZIR]

Thought it was about time to post my progress since I ran into an interesting issue.
Finished the theory part of the course and developed an MR system. The mentoring from Nick and Craig was and is simply outstanding.
Developed an mr system with a few simple rules.
Over the last 10 years the system returned 40% with a 9% drawdown. Over the last 19 years a 35% CAR with a 15% drawdown. The system has a small edge (55% right) so makes its profits from a high frequency of trading ( > 1000 trades per year). Those returns by the way included the standard IB commission rates and the universe traded is the Russell 1000.
Ran into an interesting problem. For the last 19 years the system is great. From 1990 to 1996 the results were awful with one continuous drawdown.
Talked to Nick about the problem and his observation was that markets change. He gave me three options; optimize the system from 1990-1996 and see if the system was robust, trade more than one system or discard the system. I elected to optimize the system and at the recommendation of Nick will optimize every 6 months.
I tried to figure out why the system did so poorly in 1990 to 1996. I backtested another system whose rules were close<

Meanwhile started trading my account at IB today with batchtrader. Elected to go with an exploratory live account at IB rather than papertrade
Len

[JulianCohen]

Well it’s true that markets do change. Nothing stays the same and yet at the same time the same basic principles that applied in 1900 still apply today.

When I tested some of Connors’ strategies they did really well from 1995-2005 and then tailed off so that their last few years results were single digit CAGR. This could be due to many more people using the same systems to trade (RSI overbought and oversold) so it wears away at the edge. It sounds like the system you have discovered is doing the opposite. If markets now do have shallower drawdowns, possibly because of more algorithms looking for the pullback to buy, then your system is taking advantage of that. If it works then go for it.

Trade it until the edge isn’t there any more and then look for another one.

[LEONARDZIR]

Julian, my system stays strong over the last 5 years.

[JulianCohen]

Len Zir wrote:

Julian, my system stays strong over the last 5 years.

Yes, sorry maybe I wasn’t clear enough. When I tested Connors, his systems have died off in the last five years as the markets have changed against him. In your case the markets have changed in your favour so I wouldn’t be too concerned about pre-2000’s. It is interesting to see that the reason appears to be the pullbacks being shallower.

[LEONARDZIR]

I encountered an interesting problem with IB using MOC orders and significant margin. Starting at 10 minutes before the close in the US market IB requires that you have 50% of your account value in cash or you incur a margin violation. Near as I can tell from talking to IB they have the right to start liquidating your position at the market until you have 50% of your account in cash. I asked if you have repeated violations and they said no issue but your position could be liquidated. There are 3 options. 1. Continue to trade and ignore margin violation messages. 2 Close positions at 3:49 Eastern standard time (US) at the market. I have no way to test if closing on market orders at 3:49 makes a difference in my system outcomes. 3. Trade with less margin but that dilutes my results.

[JulianCohen]

Len Zir wrote:

I encountered an interesting problem with IB using MOC orders and significant margin. Starting at 10 minutes before the close in the US market IB requires that you have 50% of your account value in cash or you incur a margin violation. Near as I can tell from talking to IB they have the right to start liquidating your position at the market until you have 50% of your account in cash. I asked if you have repeated violations and they said no issue but your position could be liquidated. There are 3 options. 1. Continue to trade and ignore margin violation messages. 2 Close positions at 3:49 Eastern standard time (US) at the market. I have no way to test if closing on market orders at 3:49 makes a difference in my system outcomes. 3. Trade with less margin but that dilutes my results.

Hmmmm…interesting. One other solution is to keep excess funds in the account. Say you are trading with 100,000 that is then margined but the account has 120,000 or 130,000 then this should be avoided. Only problem with this is that it prevents you from working the funds to the full extent.

One thing to be aware of is that should you decide to allow IB to liquidate the positions, Batch Trader will not very likely recognise this and will issue MOC orders for the liquidated position which means you end up short after the close. I had this happen to me when paper trading.

[LEONARDZIR]

Julian
Still considering closing all positions at the market at 3:49 pm. Alternatively checking in at 3:45 and see if they have issued a potential margin violation and the closing all at the market. That defeats the purpose of trying to stay away from the computer during the day

[JulianCohen]

Just want to double check Len…if you set market orders at 3:49 then isn’t that 1 minute before market close? Won’t they have already started to cut your positions by then?

[JulianCohen]

I asked IB about this and here is their reply:

The limit to what a Broker Dealer can loan a customer account at the end of the day is the “2 to 1” leverage limit as defined under Reg T rules. The Reg T account’s Special Memorandum Account (SMA) does have to be positive by 15:50 EST or the Liquidation system will close positions to ensure the account remains compliant.

However, if an account has a position that is too large under Reg T where SMA is negative and there is a Market-on-Close order entered (Not a Limit-on-Close) before 15:50 EST that will reduce the account’s position size so after the MOC trades the account’s SMA would be positive then the margin system will not Liquidate a position at 15:50 EST. MOC orders to the Nyse can not be canceled after 15:45 EST and MOC order to the Nasdaq can not be canceled after 15:50 EST.

[LEONARDZIR]

Julian
Thank you for your thoughtful reply. The American market closes at 4pm east coast time. My solution to the was to test dropping my total number of max positions by 1/2 without significantly altering results. Meanwhile you to have gotten more from IB than I have. It sounds like MOC orders will cancel the margin violation problem.i’ll double check with them.
Len

[LEONARDZIR]

Julian
Just spoke to IB. If you have MOC orders to close all positions there is no margin violation. You were correct. The person I spoke to at IB told me something different. Thank you again for your reply.
Regards
Len

[Author]

Posts