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)
How much Money for faster Optimization Backtesting: Willing to Pay.
For me at least, the only part of your 3 points I really agree with is "less than better."
"Too much decision making," "second guessing after testing" and "reasonable drawdown with in sample data" are not reasons I limit # of iterations.
But I guess the reasons do not matter as much as the final desired outcome - limiting iterations.
If I only had a nickel for every backtester who told me they needed a faster computer...
Then again, last November bought a top of line CPU to do more backtesting. I've run about 50 million unique backtests during that time. But never more than 100 iterations per strategy...
Can you help answer these questions from other members on NexusFi?
Secondarily this can cause you to loose focus on the bigger picture.
You wnt to use money wisley and hedge based on a descision, so I have en edge in one product, you can understand there are other methods to work in favor that are inversely or non directly correlated to my current edge, which is how to hedge already working captial.
Does Multicharts use multiple cores for any other operations ? For instance, it would be great if one could put multiple workspaces into their own core.
Sorry for late response. Thanks for your advice sir.
I followed your advice:
1. Decreased the amount of optimization parameters and got back faster results.
2. Cut the back test data in half. In sample and out of sample
3. Optimized or Curve fit the trading idea on in sample data to my liking and budget
4. Test on out of sample data
5. If I like the results, do some market replay and sim for no errors,
6. Go live
I have 2 algos running now and making money. Working on another algo now.
Thank you alot and others for all your help and motivation. I guess all that manually trading paid off for something and gave me some good trading ideas to back test.
Algo trading is not so bad after all and makes alot of sense. I have no issue with discretionary manual trading, it was just too hard for me, and I was never consistent making money. too hard for me. I look at charts to seek trading ideas for my liking only.
kevin how do you ensure accuracy in the strategy analyzer? I was optimizing a strategy that didn't need tick data, only needed candle close info and got wildly different results as compared to market replay for the same time period. I manually checked the market replay and it was spot on, but I don't feel comfortable with only 3 months of data on a trade
I use Tradestation, not Ninja, for strategy testing, so I can't really tell you what the issue is in your case.
I can say Tradestation is accurate when the strategy is set up correctly, and I am surprised at your comment about Ninja testing - I have heard it is good, too.
yeah it's actually quite awful. Here are the results for a super curve-fit overoptimized dog-turd strategy I was working on:
From NT:
+13,000, 244 trades, 54,000 max drawdown
Actual
+21,000, 15 trades, 2875 max drawdown
I would say what you programmed and backtested is not the same strategy as what you ran as actual. I don't know what is different, but there is probably an explanation. I'm sure one of the Ninja experts here would be happy to take a look.