NIFTY,NIFTY50,NSE,INR,1,1,FALSE Huh, well after trying this again with a different multiplier it seems to work.
Ugh I guess I missed that. So to handle this going forward, is the best option to just add NIFTY-IN as a new instrument and add NIFTY to stale? IB only has data for the latter from October 2023. Or append the new instrument data to old prices somehow (could we just change NIFTY params to match NIFTY-IN then just roll?)
"best option to just add NIFTY-IN as a new instrument and add NIFTY to stale" That's what I've done; "could we just change NIFTY params to match NIFTY-IN then just roll?)" Don't do that. Rob
I'd use daily vol for position sizing or perhaps tick data but with an EWMA equivalent to the 30 day span for daily data. And I've never used tick data except to model execution so I have no idea at what time horizon momentum will work .... my work with hourly data suggests mean reversion works better intra day and the trading costs of course are punishing. Rob
Fixed: At 1000 contracts a month for the first tier, that's many, many times what I trade with a $500k account. Rob
Semi regular update I normally do these twice a year, but I haven't got round to a mid year one yet. TFW when you're 1.5% below HWM, and checking the p&l too often... YTD: +9.5% Tax year to date: +12.8% Had an interesting coffee exactly 9 days ago with a guy that is running a similar system. He decided to 'pause' it, basically closing positions and not trading for a while. Some of this is to do with the fact he's been in a bit of an unpleasant drawdown. But also because he said he was worried about his short bond position, which was unusually large. Now as you know, I hardly ever look at my positions or risk. But sure enough when I checked I was indeed very short bonds. And sure enough the very next day there was an FT article about CTA bearishness (https://www.ft.com/content/e67a4b39-6b9d-4333-ad2d-9ce132abac9d might be paywall sorry); which also reckons that CTA positioning is predictive of yield changes in 3 months time (probably nonsense, but there you go). Of course I've done nothing about it, and I note in passing in the last 9 days I've made about 3.5% of which about two thirds was in bonds, and about half of that in US 20 years. I'm not sure what the moral of this story is, and I am the last person to draw any statistical significance from 9 days of returns, but it makes you think. My grand project for this year of refactoring psystemtrade is stalled, and I haven't really done any meaningful research. I'm busy doing a fair bit of consulting (with a crypto HF, may god forgive me), and a side project coding as part of a voluntary role that keeps me very busy. Just to make things interesting I'm currently rereading the Jim Simons book which is giving me quite the itch to do some research at some point; I keep having crazy thoughts about stat arb in futures, weird PCA factor models, and quasi TA pattern recognition... More seriously my research agenda if I ever get round to doing anything will probably involve mostly what I talked about in December: relative value models, but not explicit trading, instead pumped through DO. Creating the technology for these will help somewhat if I do start going into stat arb for some reason. some kind of execution overlay probably related to mean reversion (this is related to the research related to faster trading which I alluded to at the start of the year, which has mostly produced nothing of interest) I think I'm pretty much done with adding markets; with the joys of snap to mid I now have a bunch of markets I'm not paying data for, pushing the total up to over 250 although without duplicates I have 'just' 186. In March next year I will look at calibrating their costs given all the spread data I will have collected, calculating position limits given liquidity, and then adding them to my system which has a mere 170 markets in it now. My long term book writing plans have sort of crystallised around the following (in order): a delta one version of Leveraged Trading; basically simple system trading for dummies without access to leverage; covering stocks, ETF and (gulp) crypto. Will include factor investing, so elements of Smart portfolios. a book on backtesting and uncertainty in financial markets, quite theoretical in places but more accesible than say Lopez de Prado; potentially a good text book for my university course, also including my thoughts on managing researchers in HF A popular book which currently has the terrible working title 'secrets of the quants' where I try and explain things that I think everyone needs to know about financial uncertainty and theory. A novel (?!) But I won't be thinking about any of that until late next year at the earliest. Finally as I am sure you would all like to know, sales of the new book seem to be going okay. Until recently I had access to almost live sales figures, but my publisher has been bought out (which hopefully won't result in any changes in service) and they turned that facility off. But I'm still expecting to see my biggest royalty cheque ever in a few days time (I get them every six months), as when I last checked it had sold over 2.5k copies. For context, that's over 14 copies a day which is about twice what my first three books sold in their initial release period. But for more context, that cheque for six months of book sales is less than half what I've earned in trading profits during the same period! Rob
For me the situation is not as good - in a drawdown since last October, and not much movement in general since March, a bit up\a bit down.. but I'm "only" trading around 100 instruments and much fewer rules.. Btw, that old link seems to be down "https://fundseeder.com/trading_account/1246", looks like Fundseeder started to change something a while ago and never really finished..
Also for me is the situation not as good as for @globalarbtrader My account experienced a sudden drop in March this year, losing about one third of its value in 1~2 weeks time. It has not yet recovered from that drop.