Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
If that's the amount of data you have to go on, then personally I'd ignore it and go with the system that fit my understanding of the markets the best. Look, if I flip a coin 19 times I can easily get a "win %" around 80%. Should I trust that I'll continue to get 80% going forward? No way.
When you consider that the same goes for your average win and average loss, etc., they are essentially still unknown parameters, and the backtest data at best just gives you a warm and fuzzy feeling.
Trading platforms try to convince you that they tell you what you need to evaluate a strategy, but they absolutely do not. To name just one Statistics 101 issue: if you are taking a frequency count, and you want any chance of having a reliable prediction of future long-term results, you need that count to converge on a single level with diminishing variance over time. That's how you can verify that a coin has a 50/50 chance... even though you can get long runs of 100% heads, over time you settle on the 50% level with diminishing variance. But guess what, it can take 100's or 1000's of flips before you settle down properly on a value you can trust. Yet fancy backtest spreadsheets encourage retailers to bet their savings on 42 example trades, and no regularity analysis.
Just like in Vegas, they know that some backtest runs will get lucky and appear to win, and that customers will blame themselves or the markets when things don't work out as planned. It's a total con.
Richard - Completely understand that this backtest is not statistically significant in ANY way.
My question I guess was more of a hypothetical on what is better... Higher efficiency (less trades, less profit, but higher profit per trade) vs. Lower efficiency (more trades, more Profit, but less profit per trade).
Which one is more advantageous? In general do you give something up to achieve more profit? Obviously more trades = more risk wich I can see as an obvious downside, but other than that i'm at a bit of an impasse.
Richard - you're crazy smart and some of your statistical .pdf's have left me in awe of applying mathematics to understand situations like this and I guess thats what i'm getting at... is there a mathematical advantage to one over the other?
Ok, let's say you can trust your numbers, and that you can trust that future time periods will provide the systems with about the same number of trades as they had during the test. It really comes down to how deep your pockets are, and how much you care to risk.
If you want to make more total profit, you go with the one that made the most money. You might want to check risk of ruin and compromise, though, because often the best performing system long-term has the biggest drawdowns along the way!
On the other hand, if you want to be profitable at least x% of your days/weeks/whatever, you need to look at your profit factor versus the number of trades available to you as outlined here: How To Be Consistently Profitable in the Markets|Move the Markets (hard to believe I wrote that over 3 years ago now. ). This time, the danger is that you are profitable most of the time, but not by much. So, you starve by avoiding bad days.
Probably the best system for most people is somewhere in between, and no free lunches means you have to give up some profit for consistency, and vice-versa. Leverage on futures means you can also play games like seeing if doubling your size on system B makes it more attractive than system A, if system A makes more money with unacceptable drawdowns but system B earns less with a smooth curve.
I'm sure there are other philosophies and opinions about tuning a single system, but that's what makes sense to me.
This isn't totally true though. While more trades will obviously put more capital at risk, you reduce your risk that you are not just trading variance with more trades, especially since it will give you a much better feel if your EV+ after transaction cost.
Also why limit yourself to a or b or c or d?
What does it look like trading all 4 togather? What does it look like trading A + C + 1/2 capital/leverage with D and not trading B at all?
^ I've actually kind of done exactly that and am trading two of the four in tandem and the results seem to be the best yet. While statistically insignificant, the current backtesting results (using around 400 trades) are yielding profit factors of 5.1 and 6.0 respectively.
And the results live with them have been exactly in line with the backtest results. To say the least i'm super stoked but again, without the help of Richard and others figuring out what really matters when looking for a profitable method i'd still be spinning my wheels.
Thanks again to everyone that kind of pushed me along into the right direction!