| RESULTS FOR DAY | |
|---|---|
| CL Contracts: | 5 |
| Net $P/L: | -770 |
| Wins: | 0 |
| Losses: | 5 |
| Win%: | 0 |
| Avg$Win: | 0 |
| Avg$Loss: | -154 |
1 hour ago
Friday, April 27, 2018
Thurs./Fri. 4/26-27 -5.4%
2:20pm CDT - Well this week sucked. After 3 weeks trading CL system and the very first trade being the best winner in years, I'm averaging -$35/trade and a 10.5% win rate (backtest avg. was 19%). I knew it would be hard to trade such a system, and I'm sure I'd feel differently if I currently sat above the zero line, but this drawdown is nearly as deep as expected in an entire year. Need to do some thinking. Kick it to curb? Change from +6R system to something much less? Return to incubating? Thoughts welcome. For the week, including ES losses, net -$2306.
Subscribe to:
Post Comments (Atom)
Posts
It would seem quite a conundrum to isolate data mining from good robust system creation.
ReplyDeleteI'm curious as to what you do to avoid following the past too closely.
Not sure I follow your question. The past can repeat itself and I'd expect future drawdowns to be approx. the same magnitude. I'm either unlucky to have a similar sized max. drawdown right away or my system is flawed/failing/over-optimized.
DeleteWell, in a nutshell, how does one keep from over-optimizing?
DeleteThe million dollar question! Limited variables, walk-forward analysis, saving some data to test against optimized period, etc. Books have been written on this large topic. Still it's never easy and never a sure thing.
DeleteWhy not try a high win rate system? A high win rate signifies a particular pattern actually exists in the market, conversely, low win rate systems are trying to exploit "non market patterns". Why not focus on trying to find actual patterns, and then let the pattern dictate how many "Rs" it is worth.
ReplyDeleteYou got one for me to try Anon? The higher the Reward:Risk, the lower the win%. The lower the R:R, the higher the win%. I disagree with low win rate systems exploiting "non market patterns." Sure they could but no idea how you can draw that conclusion.
DeleteHigh win rate systems should be more reliable because they tap into structural patterns, structural patterns shouldn't change that easily over the short term if tested over a long enough period. In other words, the movement is a part of the structure. "high win rate" is relative here.
DeleteYes, as you noted, high RR systems may be profitable, but because they do not rely on actual price pattern/structures, they may degrade easier over time.
If you say so, it must be true.
DeleteUsually I try not to be too forthright with my thoughts and convictions...but you asked so here goes! :)
ReplyDeleteI'm convinced you would make money nearly every month (with comparatively small down months) if you did the following:
1) Take a very simple, logical idea.
2) Create a simple structure around it that allows the market to decide how much you make.
3) Trade it with small size and forget about the result until trade 10 (or 20/30).
I think too much back-testing is the wrong approach but it's understandable, given your background, that you'd find this a hard habit to kick. The secret is in the persistent consistency of a simple, slight edge.
P.S. I have one and would invite you to see it play out real-time but, as we discussed, that would mean you getting up in the early hours of your morning. Maybe you could adapt it to the US session?
James, you really hit the nail!
DeleteAgreed on backtesting... Besides, past performance is not indicative of future results.
If I were MBA, I'd take you up on your offer.
Thanks for 2cents James. I like simple ideas too but finding one with an edge is excruciatingly rare! Backtesting is the way, for me, to prove an edge may be present.
DeleteA few extra thoughts.
ReplyDeleteTrading an 80% losing system is equivalent to trading an 80% winning system - it is tantamount to hunting for outliers.
The market is more likely to move 4 points than it is to move 10 points, especially intraday.
Your affinity for high reward systems and your method of building a system around the numbers (vs letting the market decide) may be something to reconsider.
Good luck.
Thanks. I have no affinity to high reward systems, just ones that produce an acceptable looking equity curve over time with limited drawdown.
DeleteI'll chime in only because you asked.
ReplyDelete1. I believe technical analysis is as good as it gets for retail traders. My dull edge in being correct about market direction can be based on chart patterns, support and resistance levels, and the use of a few indicators (SMA/MACD) to stay intellectually honest with whats happening in the market.
2. Piggy backing off of Anon and James, keep it simple and back test chart patterns of your previous winners and losers, 20/20 hindsight may give you some perspective on the kind of trades you are most successful with. Being correct about the larger trend on a 5min 15min or 60min chart may help eliminate trading in chop, using a small trading size can allow for larger stop which are vital in the current market conditions.
All easier said than done.
Keep it simple , bro... best way for the retail trader to make money in this game in my humble opinion. Wishing you the best.....
ReplyDelete