The report talks about a 20/120 day crossover (not 200). That roughly translates to 1 month over 6 months of trading days. I've seen others using a similar approach, such as 1 over 6, 1 over 9, 1 over 12 months. The crossover points will differ slightly but that does not have a huge impact on the overall results.
Yes. It was a typo. You are correct. I have seen similar studies, in fact I posted this in other thread. https://www.elitetrader.com/et/thre...onsistent-strategy.324313/page-3#post-4718751 But since that study (2012), returns have not been that great.
Indeed, it was that post which pointed me towards Meb Faber's work. I read some of it and have played a bit with the system he describes. I've also seen an alternative where not the latest month is compared to a longer period, but the month before latest is being used. All these are variations on the same theme. One thing that these methods do differently though is that they only look at the close price of the month. They don't look on a rolling daily basis. The SG trend indication uses rolling daily lookback periods.
Run two threads. One monitors the connection and when there's a timeout on receive, halt the main trading thread. That's one way to go about it. Even in the cloud you need to be aware of connections drops, however unlikely.
Over 2 years later I've finally got around to investigating this https://qoppac.blogspot.com/2019/02/skew-and-trend-following.html GAT
Hi GAT, I always wondered about the purported skew of trend following. I believe that if you look at the distribution of the P&L for trades instead of the distribution of returns, you should see large positive skew (this holds assuming that you exclude rolling trades of holding long-trending positions, based on continuous futures). By the way, how is everybody doing? I imagine everyone must be printing positive numbers after recent moves (I don't trade Palladium :-D )
Hedge fund Renaissance pulls back on hunt for market trends: https://www.ft.com/content/23edf3ba-591f-11e9-9dde-7aedca0a081a Renaissance Technologies, one of the world’s most influential and secretive hedge fund firms, has sharply cut back its use of strategies that bet on patterns in futures markets, a big sign of such strategies’ waning popularity. US-based Renaissance, founded by former cold war codebreaker Jim Simons and with about $60bn in assets, reduced its use of such strategies in its Renaissance Institutional Diversified Alpha (RIDA) fund by two-thirds near the end of last year, say people familiar with the matter. The news has not previously been reported. The move comes after a long period in which hedge funds’ time-honoured strategy of following market trends has struggled to replicate past returns in markets dominated by central bank stimulus.