Search results

I run algo backtesting on the GPU for N parameter permutations and calculate a trading score for each, from which I pick the parameters for the best result. I can now quite easily write algos to run on both CPU & GPU on my platform and use the same code for both without need for separate...

OpenCL got me nice 60x performance boost with my weak 3 TFLOPs laptop GPU (vs multi-threaded CPU optimization on 8 cores). This combined with dynamic programming for walk forward optimization I got 340x speed-up (1y optimization window for 2y with monthly optimizations). This is quite nice for...

That's exactly what a robot would say ;) Just answer this one question and if you can't click the checkbox that tells you the truth:

It's in the plain sight in her name, she's running on NVIDIA K80

I found this video hilarious when this scammer realized she is criminal for scamming people and doing tax avoidance and would be f'd for being in a Youtubers video :rolleyes:

Mark says they'll plan to have only the hardware infrastructure in place to support AGI in place by the end of the year, not that they have AGI by the end of the year. Big difference. Setting up the infrastructure is the relatively easy and predictable part of the effort, and I think it'll take...

Renaissance Technologies likes to disagree

I once had a bug in my backtest code which allowed me to peek one 5sec bar into the future, which gave some incredible returns.

Minor execution detail left as an exercise to the reader :p

I have even a better strategy, buy Amazon in 1997 for $0.1 and sell it today for $155, 155000% profit babe! You can pay me $1000 for more stock tips ;)

You can pass the function object to other functions, which can then call the object without having to hard the code to use a specific function. For example, if you implement a sort() function, you could pass compare function to sort() as an argument which compares two items used by the sort()...

I'm saying the opposite :)

Yes, I have been through this exercise of generating truly random charts that look to a human as if there was a tradable signal and that look like as if it was a real chart. While we can agree that these generated charts do not have a signal doesn't mean there isn't signal in real charts just...

Exactly because it's only 90% :) If the markets were truly random this number should be very close to 100%

Strong claim from someone whose following sentence is nonsense :rolleyes: If the price movement is random there is no difference from a monkey hitting the buy & sell buttons randomly. For random price action you'll find out that over time your equity curve will tend to zero.

If random walk theory was true, then there were no consistently profitable traders and everyone who claims so is a liar.

Check "Putting the Kelly Criterion to Use": https://www.investopedia.com/articles/trading/04/091504.asp

Hi @fan27, For backtesting I just evaluate a selected trading algo for each candle. It's a simple loop that iterates through the candles and calls the algo interface. I got a bit of generic C++ template code there to make sure it's all compile-time optimized (i.e. no virtual calls), which is...

I think a part of confusion is what people mean when talking about HFT. If you talk about the lowest latency HFT then you talk about FPGA servers glued to the side of exchanges with latencies in micro- or nanoseconds. But some use the term HFT more loosely to mean "positions hold for few...

You would first need to know what inputs the trader uses for the trading decision and try to minimize that input. E.g. if the trader does naked trading purely based on price action without outside influence or if there's fundamental analysis or news that influence the trading. Then you could...