Practically every system has already been coded and is publicly available on Quantopian and QuantConnect, others are available for Ninja, Meta, Sierra, AmiBroker. You can copy and modify them as well. People who think they have a secret alpha-generating system that they never backtested are idiots. Even if you have something that has potential, it’s no different from stuff that you can already find online. And if someone wants to spend months or years improving something then they have a million of systems to choose from. While the guy who actually made more money then everyone here combined, Sarao, is the one who hired a programmer. Was he such an idiot?
Great to hear, when you say manually I assume you record your session and then watch it? Another question; how many times / trades did you have to backtest manually until it resulted in optimal set of parameters and variables?
Yes, Sarao was an idiot, or rather a criminal. His system was based on abusing the system, abusing was the system. Your oneliners are just nonsense without any proof. You don't know how much Sarao made, you don't know how much each member on ET made, so your statements are pure nonsense. You speak like a real guru. Guru's assume that people believe anything the guru tells, just because he is a guru. So you can give me then the system used by Renaissance, Virtu and all the others? You even have no clue what exists in terms of systems, and you are not able to check the whole internet to check what systems there are. How do you know that all the systems are on internet? Can you send me a copy of my system? It should be on the internet according to you. If not we have the ultimate proof that you are just telling BS. If everything could be found on internet there would be millions of rich traders. Reality however is that there are millions of poor traders. You are a millionaire as you found a supersystem on the internet?
See, even you cannot agree that he was an idiot. I didn’t say that highly sophisticated systems that won’t work with basic software are available online. But ok, your logic must be flawless so I’ll stop right here.
It is pure logic that if you make a statement, you can proof it. You cannot proof any of your statements. You even tell now the opposite of what your statement was. Highly sophisticated systems are the only systems that perform very well in any circumstances. So these are the exactly the system you will find nowhere. There is a huge difference between flawless logic and the statements you wrote. Logic should be as close as possible to flawless, if not it have no value.
Look, the OP simply asked how to backtest a strategy without programming. So he thinks about programming but for now can’t do this. So logically one of possible answers is to hire a programmer. And there may be ways to do this without disclosing every piece of information. While he doesn’t need to do that either, and many other answers can also be valid. He simply doesn’t need to use my answer, but may consider it. I’ve had multiple programming teams working for me in the past (non-trading projects). While I’ve also hired math professors to help with some trading tasks. There are many solutions to every problem.
I guess I am an idiot then, as I have something that my competitors do not have and it actually generates alpha in real life. Same goes for PDT, the ETL team, RenTech and many other successful funds or PMs. There are infinite dimensions to alpha generation and ability to come up with new strategies comes down to a combination of creativity, knowledge of the market and the skillset to actually trade it. It's a pretty rare skillset, believe it or not. Try googling and you not gonna find anything about real life intraday statistical arbitrage, high frequency trading or alternative data strategies. For that matter, pretty much anything that is being currently used in the industry is not in the public domain. People who do develop good ways to extract alpha guard their secrets in all ways possible. For example, 1-year non-compete clauses are standard and I am hearing of 2-year non-competes more and more. People who tried copying strategies or even benign bits of code have gone to jail. So there are a lot of real secrets out there and no, you can't find everything online. Most PMs/fund teams have multiple people, that's the only way to gain leverage. There are two possible approaches to alpha leakage in these cases. One approach is to build a cohesive team that is very loyal to each other and sticks together. While there is usually a clear leader, the team shares everything and trusts each other explicitly. It's a highly collaborative approach that yields the best results, in my experience - in fact, all successful teams I've seen where built on that approach. An obvious drawback is that hiring new people is very hard, as your choices are very important. Others have instead built a highly compartmentalized process where their employees have only partial knowledge of what's going on. In this case, programmers and quants are replaceable and there is relatively little risk in them leaving. Of course, this makes for a very inefficient process where no single employee has a big picture view. In Navs case, he has been with that programmer for years from what I understand and was sharing a lot of information with him. While it was working well, ultimately this was also his downfall - the email exchanges was the key evidence in his trial.
Don't Rentech and most funds backtest their strategies, especially those that can be automated? Though sure, people traded before backtesting was feasible, thus there are people with experience, education, and capabilities to trade based on information. The conversation here was started by OP asking about backtesting, while I think my response is valid for any new traders. And my response was to someone calling me an idiot. Though I was actually an idiot by thinking I figured out how to trade UVXY options before I could backtest it, so there is plenty of idiocy going around. My responses here were also in context of people new to trading whose strategies are really no different than any retail strategies already published, sometimes sold, or at least tested on platforms like Quantopian, which as we know failed with all their 100,000+ retail quants, traders, and people who test their millions of strategies. Any people capable of generating serious alpha probably wouldn't be posting questions here, or at least not in the context I've seen discussed in this thread. While I see plenty of people hiring developers to backtest strategies, often they may not have a choice and if their strategy is not highly technical/clever/unique or based on unique sources/information then they indeed should backtest it any way they can, including hiring someone, if can't do this themselves...
Oh, I guess I missed most of the story, sorry about that I read your message as "everything is out there in public domain, its not worth doing original research and there is no such thing as secret alpha". Mea culpa. My personal approach is that backtesting is one of many ways to glimpse into the merits/drawbacks of a strategy. It should be done but it is not always possible or produces realistic results. Instead, one can also build strategies based on fair value models, based on similar strategies that are possible to backtest backtests etc.