- This topic is empty.
-
My wins/shortlist of points on MOC are:
A. Trade smaller cap stocks. I’m talking R1000 and R 2000.
B. Trade them separately. Don’t just do one system/settings for the whole of the R3000.
C. Keep minimum price double digits. Below 10 USD and you have to buy really large quantities.
D. Along with point C, ensure plenty of positions such that individual position size is small. I’m talking to make sure that any single LMT order is no more than about a few thousand units at most. Any more than about 5,000 units in one order is really hard to fill in a snap second when the price drops to your limit. If you are doing a $2 minimum on a $20,000 position you would need to buy 10,000 units in a snap second which I have found to be damn near impossible with IB. If you are wanting 1,200 pieces of a $20 minimum then your $24,000 position has a much better likelihood of filling rapidly when your price hits.
E. Be careful with rounding and tick size of prices used in explorations/backtesting. I have recently forced all my trades/exploration in this system to a 1 US cent rounddown using Prec(BuyLim,2) and this has vastly improved my backtest results compared to real trade (you may have seen some posts in the forum, especially Julian’s thread, where we’ve discussed backtest results looking good, but real time trade showing missed trades whenever the limit price is right on the cent or within 1 cent).
F: Keep Entry criteria pretty broad. Don’t try to get too tricky with too many conditions/indicators for entry otherwise system will be too curve fitted/niche for certain conditions and you won’t get many orders/fills.
G: Make sure the profit distribution graph shows “two tall buildings” in the skyline. That is, -5% and +5% number of trades more than 10,000 on a ten year or more backtest, with the +/- tails in the single digits or thereabouts. You wan’t your profits to be in the guts of your trade volume. If your system depends on only making profit on a small handful of long tail winners, you know what’s going to happen, that will be the day you are unable to enter your trades because you were busy etc.
I trade R1000 and R2000 with 90 positions each. This week for example, on Monday my explorations had over 1000 trades (I obviously only use the top 90 + 90 when entering the orders) and after the session ran I had 109 fills. On a regular, non-ball tearing day I usually like to see 10 to 20 fills. On a slow day I might get 5 or 6 fills.
For systems like these where trade volume is the key, I try and aim for a minimum of 1000 trades per year in the backtests. If we are to follow Nicks “Next 1000 trades”, the sooner you can get through those 1000 trades so you can see your edge play out, the better. If it takes three years to get through 1000 trades on a system like this you are probably going to throw in the towel early.
- You must be logged in to reply to this topic.