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)
Few reasons:
1. Summer - much less opportunities, opportunities that present themselves are mediocre mostly and result overall in around breakeven results, because moves are small, and slippage is present and account for significant part of the move, because of low liquidity and participation
2. I want to capture more ticks with less staring at screen, hence going for more swingy approach, plus it means I can utilize my allocation of buying power better, as I can get better return in $ per each contract used
Hi Xelaar, I just have a quick question about how you test your new setups. It seems to me that for you the time between coming up with an idea and actually implementing the new setup to your live trading is fairly short. How long does it usually take you to start trading a new idea. I'm curious how the usual data collection phase looks like? How big sample size are you looking at before moving to sim and eventually putting real money on the line?
I'm interested in the testing process for short term trading ideas as I'm testing a couple of setups myself right now and it seems to be a never ending process for me as a I keep on thinking that I don't have enough data yet.
PS! I have silently observed your TST adventures and I'm very happy that you have decided to keep on documenting your everyday trading after the combine. Keep it up!
Well I can only guess how you decided the time between introducing an idea and taking it live is short. But it certainly not so. Here is a thing with testing and doing all data mining for any short term trading ideas: if it takes you many months to get through this process before taking it live, you can effectively forget it. Most of short-term trading ideas are based on market inefficiencies, that present themselves most of the time, but their period of existence is fairly short: from months to days. That is because many HFT firms are also hunting for them, and it takes them much faster to do all the data mining and take it live.
Another thing is a lengthy back testing. It makes no sense in my humble opinion just because market changes all the time and these inefficiencies are only present within a certain time window. All backtesting will give you is that over the period of years you would lose you all account. Shorter the time frame of potential trade, shorter the back test period should be.
It presents a dilemma how can you judge when you have to stop and move on to the next idea, instead of trying to persevere and weather the storm. I think this is a hardest part and one either can have a mathematical answer or a gut-feeling answer. I have been through many of these ideas and while not all were successful in the end, overall I was very profitable even as many people lost money with them, as they trying to "persevere" far too long and gave back all the profits and much more.
Market has a certain feel. This is the beauty of short term charts and watching price action and order flow unfolding. You feel the beat. And as any doctor can notice when patient's heart beat starts to go wrong, attentive trader can get a good feeling market is changing as well. If same is true for longer term traders, they might notice it unfolding through months or even several years, for short term trader this may happen during the day and more than once in most extreme cases.
So to answer to you - I don't have a really formal process. I look for charts for last few weeks, notice an opportunity, mark failed and succeeded setups, pen overall results, scan some more charts and look to do forward testing for 2-3 weeks before taking it live with very small risk. The problem of very long testing is that by the time you complete it and move live the setup might already expired and you will never get out of drawdown. My approach is to start live trading sooner but only with a small risk and build up any increase of risk based on profits already collected.
To add a bit: I see many people are entering markets with academic or engineering backgrounds (I have one too) looking to excel in it using their scientific prowess. By doing so they mostly are looking to find certainty and something stable they can rely on forever, or at least for years to come. This is the biggest fault in my opinion: there is no certainty in markets, there is no stability in trading and there will be no consistent income from trading. One month you can make a killing, another month you will crawl on you belling home crying. If people want to make it in trading they have to realize it and accept it. I think books of Nathan Thaleb are pretty good to give some good clue about the randomness of this whole thing.
I see thorough backtesting only as a search of consistency that just isn't there. Not saying it worthless, or no good system can come out from it. Just that it won't bring that consistency one is looking for. Markets change. Systems, setups and trading must do too.
That's why no any mechanical/automatic system ever made huge success in trading and took all the market's money. Great algo house we hear about like Medallion Fund from Renaissance Tech have hundreds of people constantly looking for inefficiencies and adjusting countless algos for the changing markets, retiring outdated and bringing on new. I personally welcome automation a lot, but I don't believe in any algo that can work efficiently for any long term without very deep and constant human intervention.
Xelaar, thanks for a thorough reply. This definitely provides me with a new point of view as I've always looked at data mining and backtesting from the so called academic or engineering side trying to find a setup that I could rely on for years. Definitely something for me to ponder about with my tests.
I think mechanical trading is fading away in general, as smarter and smarter universal algos will remove inefficiencies with the speed of light. This process will of course diminish any returns for these algo firms but also render short term trading as more useless. I think that discretionary trading will be more and more on the rise, taking into account more and more variables including fundamental, news, technical, etc
You may notice few changes. As I am working on a different setup now, I don't need these range bar charts, and can have instead two traditional charts on screen. I also work with a larger risk, up to 25 ticks. I was given approval to use a strategy with a larger risk as well by TST.
Hey Xelaar, just wanted to let you know I also work with around 25 ticks on GC, depending on volatility. I rarely take a full bar loss though as I'll pull the trade if I think I am wrong. Best of trading to you on this next development.
