OK I just got this (no email interestingly, but alert when i logged in). So it's universal and probably IB driven. Rob
I don't allow negative correlations when computing IDM. My IDM in production is 3.0 which is capped, but in practice it's probably 3.5 without the cap. I'd rather undershoot my vol target than gear up a lot, but that's your choice. More here https://qoppac.blogspot.com/2023/03/i-got-more-than-99-instruments-in-my.html Rob
I had forgotten to mention that: I was not alerted by email about this, but was led to this questionnaire when I logged into my online account. Thus same as your experience.
Hi Rob, I have a question/comment/suggestion I'm not sure if it is interesting for TTU. It is regarding to how to use carry in your strategies when you roll from a contract to the next one. Imagine you are trading a contract where you hold a contract that is not the nearest and there is enough data to compute carry for different maturities. When you roll, should the raw carry of the previous held contract be used in the computation of the current forecast? I mean, the values that you input in the Ema smoothing should only refer to the currently held contract (when this is possible to compute). Depending on the shape of the curve carrys of different contracts can be very different between them and the EMA takes time to adapt. I have been playing with introducing "carry rollovers" in the time series of some markets to avoid this effect. I think this is different issue from the ones discussed on the chapter 'accurate carry' of AFTS. Thanks, Luis
I bit the bullet and started adding CSI data a few days ago. In the first pass, I backfilled instruments I already have and added some delisted instruments that have a reasonable history (such as pork bellies and the German Deutschmark bonds). This exercise produced some interesting results: The Sharpe ratio of the backtest increased by 0.1. This effect is due to greater diversification in the earlier parts of the backtest (70s to mid-90s), as there are now more instruments traded (I use a dynamic IDM). The next step is to add more instruments. I have identified another 111 instruments that fit my liquidity constraints. As many of you have done this work, I have to admit that adding instruments is pure slavery. It is highly manual and requires constant concentration, juggling websites, databases and Excel spreadsheets. I feel as if I had signed up for my first junior position again. Now to my question: I have come across instruments on the NSE (Nifty and NiftyBank). Are these instruments tradable by non-indian-residents? I have seen some references to an application form on the IB website but do not know what to make of it. On the operational side, I am struggling with instruments quoted in INR. As my base currency is EUR, I need a EUR.INR forex rate but cannot find it on IB. Does anyone have any experience trading NIFTY? As always, thanks for your consideration and stay tuned for how much sharpe increases going from a 100 to about 200 instruments (I guess Rob could spoiler the answer )
Pure slavery: hah, it is tedious yes, but on the upside you only have to do it once. No you can't trade INR FX which means you can't trade the futures as there is no way of getting margin in and out of the exchange. In expectation it will go up by 41.4% eg from 1 to 1.414 but that's expectation and assumes you have 200 instruments for your entire backtest. Rob
I expect to see an increase of about 0.2 sharpe (16%) in the backtest and perhaps 0.1 in live trading. Pure speculation of course
A thought. On idm's I assume the main purpose behind the cap is to take a haircut on the theoretical max correlation benefits (even after flooring correlations at zero) for reasons not dissimilar like not taking full kelly seriously? If this is the purpose, one thought I had is modify the framwork and reduce the idm used in each walkforward period of the backtest to 90% or 95% of the theoretical idm for that period rather than fix it 2.5 (or 3) through the whole backtest history. 2.5 cap is 20% higher than the theoretical idm in the early years and is never used and about 20% below the theoretical limit in recent years when there is a much wider universe of instruments.
That isn't how I think of the IDM cap, for me it's a limit on peak leverage. (You could argue this is better dealt with by the dynamic leverage limit used in the risk overlay, and you'd be right.) Also the IDM isn't measured on eg realised Sharpe Ratio from diversification benefits, rather an idealised version which assumes that every instrument has equal expected Sharpe Ratio, so I don't really I'm actually only gently capped on IDM now, I no longer stick to 2.5 and in my production system I'm using 3.0 And I'm personally so far from full Kelly it's irrelevant (25% vol target, around 20% realised, versus between 80% and 140% on full Kelly depending on whether I use realised live returns or backtested) Conclusion - release the IDM limit and use a risk overlay. Rob