QUOTE(xuzen @ May 29 2017, 11:15 AM)
Dasecret,
» Click to show Spoiler - click again to hide... «
Hmm.... what kind of monte carlo method you use to be able to come out with a decision of allocation for specific funds? I thought monte carlo helps with probability, sensitivity, scenarios etc. that's why the FSM one would just show you where the different scenarios would end up in that chart
I was just jesting with you, I am no good with actual Monte - Carlo. However, I did use various UTFs and various Correlation - Coefficient to run multiple scenarios. Then I choose the most optimal risk to reward portfolio. The most tedious part for my Algozen is data entry.
RHB AIF and Ponzi 2 have very high correlation to AMReits and ManuReits. 90% correlated. When the UTFs are very correlated, it would be wise to choose the one which offer the better risk to reward ratio. With regards to ManuReits and AMReits, these two are very similar in terms of risk adjusted performance that is to say, ManuReits give better return but it comes with greated volatility compared to AMReits. In the end, I decided to go with AMReits because it has lower volatility (more stable) and let my India and TA-GTF act as the alpha - maker (forward striker in football parlance)
However, if one wishes to substitute ManuReits with AmReits, it is perfectly OK as both Manu and AMReits have around 95% correlation. If one looking at individual UTF, then one will say ManuReits make sense as it return is better. But when one is constructing a portfolio, many other parameters comes into play.
Xuzen.
Just find it interesting that ponzi 2 which is a pure EQ fund to be highly correlated to a pure REITs fund. AIF I can understand since it does invest quite a bit into REITs on top of FI and EQ
QUOTE(Steven7 @ May 29 2017, 11:19 AM)
Well don't get too carried away with all these AI hype, speaking from a tenured Software Engineer who has worked with multiple AI and machine learning problem. Here is a good read for you
https://backchannel.com/the-myth-of-a-super...ai-59282b686c62 I still prefer a little human touch when it comes to investing and I definitely don't trust the newcomers (Smartly, StashAway) AI yet as they are nowhere close to AlphaGo, the main reason being Google is way too ahead in this AI realm
QUOTE(Steven7 @ May 29 2017, 11:26 AM)
To build on your reply, yes most AI is still built by humans with some exceptions, for instance, Google's AI is building a better AI compared to its engineers but I agree with you that we can't rely on full robo-advisory yet for now (on my previous reply). Speaking of which, predictive analytics is always an interesting problem in AI (in fact I am working something like that right now but on a different context).
Quick unrelated fun read: There is an open-source project that does the following, monitor Trump's feed, process each tweet by extracting the company name mentioned & perform sentiment analysis on the text, if its good sentiment buy the stock of the company thru TradeKing API immediately and vice versa. The simulated portfolios earns a shit ton of money.
Now, the prices of the stocks mentioned in Trump's tweet going up and down; isn't that also due to the investors manual speculation? So machines are just doing it faster to gain that competitive edgewhile the investors are still sleeping or haven't read the news
This is the kind of conversation I find interesting
Great that you are in the front line of the technological change, I know little of these stuffs other than what my brother feeds me sometimes. Since I'm in the no.1 profession to be made redundant with technological advances, I'm keeping a close eye on the tech developments and hopefully can put myself in a position to still remain relevant for the rest of my working life