Search results

Perpetual motion machine for a few weeks more maybe.

Rather like, people/machines on the sidelines being dragged in to not be further embarrassed about underperforming.

This ended up being larger than intended and is off topic, but posting anyway: A third (very common) category: Faker. Pretending to do well because psychologically protective reasons or being there to sell a service (the malevolent kind). For the former, 1 + 1 = 3 if you follow their logic over...

His profile says Pakistan so I think his English should be excused. However, I'm struggling to understand if this is a question or if he is trying to express a view.

Just my own biased opinion: "Breakouts" at sub-hour intraday chart is not worth bothering with unless you are some kind of natural trading genius, and even then you're probably not thinking about textbook breakouts. Start of day was awesome, no point trading the rest. Make things simple by...

My (completely shallow) understanding is, with options, you can buy a put to protect you against price drops above a certain magnitude, or similarly buy a call to cap loss on a short position. Being long options don't get you stopped out (margin issues aside). Of course, buying this insurance...

The alternative to using SLs is to have a small enough position that losing 100% doesn't materially impact your capital (if long, if short it's not really viable without hedging or a 100% stop on something like an index unlikely to ever gap more than that). In that case, you will stay in the...

Ah yeah, I've intended to look into AlgoTerminal as well since it looked very nice on a cursory view, thanks for the heads up about the name change.

I assumed he was sarcastic.

He's tacitly admitting to trading without any edge, my guess anyway. Some confusion can stem regarding backtesting of a strict mathematical/"mechanical" rule versus manual backtesting by a discretionary or semi-discretionary trader on events as they happen in a recorded stream. I didn't realize...

This is the most distilled (and hence best) argument. If a strategy doesn't work in backtesting (when represented properly), it sure as hell won't work in forward testing or live trading. You can rule out a lot of BS this way. The reason backtesting usually fails is A) because people test more...

It's a bit different between research (finding an edge/alpha) and execution (a brokerage like IB or Alpaca). Generally, I would say the former is necessary unless you want to be one of those idiots who "trade" but lose money (yeah it's easy to belong to that category). For the former, try...

TBH those apps suck anyway, they're doing their population a service.

DCA (and by extension B&H) are great if you live long enough and start early enough. Otherwise, it's a bit of a lottery that you will be sitting through a high stock returns macro period. DCA is still a very reasonable way to deal with investment in face of uncertainty, especially since it takes...

I mean, you can code up an entry, then backtest it with random exits. You can also code a monkey test/random entries and exits. In fact, I wrote an algo that internally compared some fixed strategies against multiple random strategies and only traded them while they did significantly better...

Hilarious thread

Yeah, QC is only back testing (and live trading or selling signals/alpha). Depends on what you're doing. Optimization is slightly 'magic' anyway and is a large source of curve fitting. On the other hand, a lot of traders do use it in practice. Anyway, if WFO is absolutely critical to what you're...

IMO QuantConnect is a decent introductory platform (Python or C#). Once you used that a while you may figure out things that you need and it cannot offer. This platform has no optimization though (which means WFO isn't necessary) since it's built for hypothesis driven strategy development...

I would recommend keeping the NaNs in there and then do filtering/mapping just before passing into the consumer function. It's likely that different consumers need different interpretations of missing data. Of course, in very performance intensive code this might not be tolerable.

Only 200x leverage, OK ;-)