Timeline for What kind of basic framework or application do you use to run your trading algorithms?
Current License: CC BY-SA 2.5
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 17, 2015 at 17:21 | vote | accept | Sebo | ||
| Apr 11, 2012 at 10:33 | comment | added | Konsta | Hm, yesterday I saw a post where somebody explained his/her execution backtest. After estimating expected latency, she chose a range of likely prices; now you can look at the average price you'd get or take the worst price in that range. I believe looking at both cases makes execution more robust. | |
| Jun 14, 2011 at 15:32 | vote | accept | Sebo | ||
| Jun 14, 2011 at 15:32 | |||||
| Jun 14, 2011 at 15:32 | vote | accept | Sebo | ||
| Jun 14, 2011 at 15:32 | |||||
| Feb 8, 2011 at 16:22 | comment | added | Pete | Imran, if I am backtesting using OHLC bars, and find an entry in bar $n$, I will generally take my backtest entry at the open of the next bar, $n+1$. This leaves some room for improvement during execution. I have also found if during execution, I do a market order entry at the start of the next bar, the actual entries are evenly distributed around the entry bar "open" price when I later look at historical data. So my point is: slippage adjustment is a function of the way you backtest. | |
| Feb 8, 2011 at 14:42 | comment | added | user256 | Hi Pete. What do you exactly mean by "...,I can actually count on getting better entries, on average, than the backtest.". I would also expect to take off some slippage ticks. In which case can you avoid that? | |
| Feb 4, 2011 at 16:47 | history | made wiki | Post Made Community Wiki by Robert Cartaino | ||
| Feb 3, 2011 at 17:00 | history | edited | Pete | CC BY-SA 2.5 |
added Edit: par
|
| Feb 3, 2011 at 16:51 | comment | added | Pete | I use Matplotlib (matplotlib.sourceforge.net), a Python library, for most visualization. The exception is when using JMP for data analysis, it has its own plots. JMP is a commercial product that I happen to have a license for through my company, and use it for convenience. There are plenty of numpy/scipy/etc. packages available for Python to do the same stuff. In fact I code stuff up in Python once it becomes part of a process. | |
| Feb 3, 2011 at 15:05 | comment | added | Sebo | @Pete What kind of visualization do you use, to better understand what your script is doing? Do you generate a R script end render it afterwards with R? | |
| Feb 2, 2011 at 19:14 | comment | added | Pete | PostGreSQL. Not sure why I use it over MySQL. Both are freely available. | |
| Feb 2, 2011 at 17:12 | comment | added | Sebo | Thanks Pete, I'm actually also interesting in doing it myself later on. What kind of database system are you using? Maybe there is a even a specialized version of a standard dbs available. | |
| Feb 2, 2011 at 16:46 | history | answered | Pete | CC BY-SA 2.5 |