GAT/Quiet1 thanks - I have dug a bit deeper and if you check the ICE website, short stg does indeed have 0.005 tick value (as both of you have observed). When I retrieve the IB API 'minTick' attr for short stg I instead get 0.01. So I am guessing they probably need to update this attr in order to correct the problem.
Sooo. I was in the process of utilizing the pysystemtrade sysdata model to collect historical data from Quandl, store in Arctic etc etc, and I *think* that all of the CME data has just disappeared from the Quandl platform. I guess I shouldn't be surprised given the acquisition by a rival exchange. Does anyone else use Quandl as a datasource, or does everyone steer clear of free data sources and build history using their broker feeds / paid sources?
How did you build your history of data? I have not explored the IB API too deeply but as far as I can make out the history call reqHistoricalData stops returning data after the contract expires. Maybe I am doing something wrong since the docs say "older than two years counting from the future's expiration date". Either way that's not much history.
You are correct in that IB only provides data for contracts which expired within the last two years (plus of course those that haven't expired yet). I only use the data for running the trading program, I don't use it to do longer term backtesting. For my running program do I need a little more than one year of historical data (I use daily close prices).
To report back, after I raised this issue with them (repeatedly) it appears shortstg prices finally have the correct mintick attr. I imagine IB had to amend a config file somewhere. I must confess I only found out when shortstg prices were suddenly being returned with an extra dp recently...not because they told me it had been fixed. I am grateful, though.
https://www.barrons.com/articles/wi...5845431?mod=hp_BRIEFLIST_1&mod=article_inline Winton Capital’s Idiosyncratic Founder on Artificial Intelligence and Statistical Fallacies Most of it is old news. Some highlights. "Winton researchers once analyzed whether traders could make money by knowing six months in advance the figure for gross domestic product. While Winton did find that traders could profit if they knew the GDP figure a day in advance, essentially trading on inside information, it also discovered that knowing six months ahead wasn’t much help. (The firm did not use the results of the study to build a trading system.) “If you can’t make any money out of knowing it, how can it be so fundamental?” Harding asks. There is probably some information where advance knowledge can be used to turn a profit, he allows. But, he says, “I have always been very skeptical about economic statistics,” pointing to the fact that they are often revised after release." ... "In an interview spanning more than two hours, Harding eagerly discussed artificial intelligence, which he thinks is overhyped. Deep neural networks, which use advanced mathematical modeling to process data in complex ways, have “some definite applications,” Harding asserts. But “the idea that you feed in lots of data about financial markets, and then it tells you what is going to happen, is for the birds.” ... "Momentum investing is one area that has served Winton well, propelling the firm to profits in the aftermath of the 2016 Brexit vote. Harding was drawn to momentum investing after observing the success of empiricism over theory in the sciences. “People like ourselves and our fellow travelers don’t look at individual situations,” he says. “We look at patterns across time and across markets.” Winton has a room devoted solely to charts that display the long-term price movement of 80 different assets from sugar to stocks. Sometimes, Winton’s systematic trading strategies can mean that its funds stay at the party even after it’s over. In February, for example, Winton’s funds lost 5% in a single month after riding a long bull run in stocks. Winton’s positions in directional stock-index futures had driven returns for more than a year, but trading turned volatile. Funds like Winton’s typically don’t do well when there is a sharp reversal in markets. Around the time of the loss, Harding declared that he was accelerating Winton’s retrenchment from trend-following, a strategy that had worked fabulously well for the firm for decades. Like most winning strategies, though, the trend-following space had become overcrowded. By next month, Winton will have reduced its allocation to trend-following by a quarter in its multistrategy fund, with $8.5 billion in assets, and in other funds managing an additional $3.5 billion. Critics of Winton still question whether the firm’s returns are sustainable, especially given that the hedge fund’s performance has been driven by trend-following strategies that are now too popular. When stocks hit the skids this past October, Winton was better positioned for the drop in equity prices. But the reversal of long-term downward trends in sugar and coffee, among other commodity movements, weighed on returns. In dollar terms, its Winton fund, which has produced an annual return of 11.92% since its inception in 1997, lost 2.54% in October. In November, it gained 1.36%."
I was wondering if anyone here has used a dual setup on their machine running both a live account and a paper account? I am considering setting one up simply to be able to compare the outcome of live trading with simulated trading but its a big process and requires duplicate dbs and careful consideration of config settings which clearly separate between the two, plus its hard to test without risking the live set up being broken by the paper account. I guess I am curious if anyone has done this and found meaningful difference between the two (mainly does paper trading overstate the performance as it might be perhaps be too optimistic about execution and performance)? While I am here, I want to thank everyone in this thread, and above all GAT. Setting up a fully automated trading system at home in your spare time is a lonely journey at the best of times, and a thread like this, as well as the many blogs from Robert, has made that journey certainly for me, and I am imagine others, a lot more tolerable. So, thanks and very Happy New Year!
It is a big process (and the other gotcha is the pain of running multiple IB logins) ... and the question is what value will you get from it? You're very welcome. In other news, if anyone is interested (I don't track it myself, preferring the tax year measure of performance) here are my scores for 2018: I lost 3.1%, of which about 2.1% was from futures trading, and the rest from my stock + hedge. It's my first loss for a calendar year, which is irritating but was bound to happen. I'm up 4.5% since the end of the tax year, so hopefully I will maintain my record of not losing money in any tax years (in aggregate; I've lost money trading futures in the last 2 tax years). Incidentally in just futures trading I'm up over 7% in this tax year. Final numbers aren't in for most funds, but AHL diversity GBP was down 2.2% at year end (this has a lower vol target, so I'd be outperforming here), SG CTA index was down 3% (again much lower vol than me), Barclay CTA down 3.1%, and other numbers are much worse; Aspect down ~10%, ISAM down ~20%. Winton appear to have made a little money ~+2%, but as is well known their programme hardly contains any trend following nowadays. Another data point is that global equities were down 12%. Is losing less than equities diversification? Anyway my 2018 wasn't great - I guess it was the same for most trend followers, although I'm not a pure trend follower so using this as an excuse is a bit lame. Here's hoping 2019 will be better. GAT PS I'm flat USDJPY so wasn't affected this morning. Thanks for asking