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)
Can you elaborate on this log returns issue, why do you think that backtests on platforms are misrepresenting the backtests results. How about a backtests with a couple of instruments?
No.. quants are not using price charts for analysis, they do things like factor analysis using a lot more data on a lot more intruments. They will look at distributions and to what extent factors are predictive. They use price data (never enough) and combine it with other data to disprove some hypothesis about a price anomali.
This is a total other ball game has nothing to do with charts or TA. I mean the indicator stuff in chart trading which can be objective but isn't quant stuff.
I watch market behavior. If the price setup and the technicals, etc. all look bearish, then it's a matter of trying to get in short. However, IF the price does not drop as intended, then that means 1 of 2 scenarios: 1) It should drop soon. OR 2) If the price setup and the technicals flip to bullish, that could spell a big move UP.
All I care about is how is the market reacting to this setup -- and that, to me, is highly predictive and actionable.
Almost nobody uses natural log returns, I didn't either until the middle of last year and I've been doing this kind of work for many decades - first strategy in the early 80s plus a lot of studying around the time listed options were introduced in the early to mid 70s.
"Vanilla" returns (End Price / Start Price) - 1 are complicated to deal with computationally. A 100% gain and a 50% lost should put you back where you started but the arithmetic doesn't work: 100% - 50% = 50 not zero. Log returns allow the logic to be mostly arithmetic instead of math.
At that time I was also producing a trade listing. A trade listing encourages you to use vanilla returns, like you invest $10K every time or buy 1 contract or whatever, so you want to know what you made or lost per trade. This gives you a fixed profit per trade.
The elimination of the trade listing was a major achievement in getting my application to work properly. I go directly from price history to performance analysis, maybe it is useless but I doubt many other developers are capable of that.
The backtesting issue is more or less the same problem, that gets caught up in producing a trade listing because the developer/coder is weak, not because you need it for something.
It is a lot easier to understand if you have actually faced the technical logic/computational issues.
I worked several months on this problem full time. A lot of it was spent just staring at the data or into space. A lot of the stuff I do is subconscious, one of the reasons I write the articles is a way of trying to understand what I am doing.
So yeah compounding geometric across time, because of the volatility drag, and arithmetic across instruments. What about your claim that this isn't implemented in most platforms, why make these claims?
If I could give you an advice is: be a bit more open for discussion, be more humble, if you can learn us something please share, also the other way around, learn from guys here that are good traders, even if they don't share the same passion as jerking off on academic topics.
I haven't seen it anywhere. I know tradestation quite well and they produce a trade list. You can also only analyze one security at a time. If a strategy works on one equity it should probably work on all or most equities, they don't consider groups at all. The literature I've read don't mention real issues much, let me know if there is a good solution not that I'll drop 30 years of work and give somebody any money at this point.
If you work at a firm with multiple traders, adminstrators or whatever there will always be some jerk who wants to see the vanilla return because they can't deal with a 100% gain looking like 69 as a natural log.
From MachoTigerTrader
Which of those qualities you list is a negative if you are serious about beating the market?
Previously I was just posting pure data with a little commentary and was curious why nobody made comments.
Unfortunately I work best with a little controversy and emotion, hope nobody takes it too personally.
Reread my post because that's not what I said. I said that a quantitative strategy involves specific calculations and output the same result given the same set of inputs every time. I did not say that a quantitative strategy is technical analysis. A quantitative strategy may use some inputs generated from the market like the price of an asset, the risk free rate of return, or volatility. However, it is also likely to include inputs not generated from the market like the security holdings of a company, the total outstanding supply of a commodity, or economic numbers. Such a strategy could use technical or fundamental analysis or both, and it could use any number of complicated methods to calculate its result. The key factor is that the result is deterministic. There is not subjective decision making involved. It makes the calculations and enters the trades based on the rules.
