Created Model with Decent Returns...Seeking Advice!
looking to get some advice on a model I've created that seems to be a bit different from the other ones in previous posts.
im an analyst with a few years of experience at a long only fundamental based shop but I've built a model that is purely quant in nature and wouldn't be given the time of day at my current shop. it is NOT a market trading model but instead, it ranks every company in a defined universe and then a portfolio of X names are created depending on how concentrated one wants to be after the companies have been run through a series of algorithms.
briefly without going into too much boring detail here are some of the highlights from the original portfolio and backtesting since 2002 (furthest back i can go due to my data subscription). i also have been watching it on a daily basis to make to make sure the numbers weren't lying...
- beats the index on a quarterly basis 70%+ of the time
- beats the index on a trailing 1 year quarter ending basis 85%+
- beats the index on a trailing 3 year quarter ending basis 90%+
- beats the index on a trailing 5 year quarter ending basis 100%
- through october 31, total excess return of 230%, annualized excess return of 13% (round numbers)
i have been considering looking for a more quantitative shop but also have toyed with the idea of opening a firm since the underlying algorithms can be translated to different market cap ranges thus different products. so my questions are...
what kind of shops would be the best to potentially approach?
is it even worth trying to approach a shop since I'm rather young and might not be believed?
if i was to open a shop, does anyone know experienced seed investors?
and lastly...wanna invest? haha
A couple of questions (that I don't need to know the answer to so don't post them to shut me up):
1) You haven't quantified beats the index X% of the time. Say I beat the index 99% of the time by 1%, but on other times i'm behind by 90% this is a signficant factual omission. If you beat it by 99% of the time by 1% and are 1% behind on the others, then you should clarify this.
2) Does it scale? I'm sure there are many models possible that could make a 1000% return on $100. Doesn't mean it will make 100% on 100,000 dollars. Again, it should scale, but also be prepared to state the range of this scalability. Not having an answer to this will make you look stupid.
3) What trends does it rely on (if any), really think thoroughly about this. If the market crashed tomorrow, are you fucked? Assuming the next day will be like the last is very thanksgiving turkey effect (see Taleb) and being brutally honest about this will save you a lot of burning later on. I'm relatively confident that some market events could break it (short squeeze, hyper inflation of a major currency etc. but you should have these thought of and explained, as well as ways to hedge against them). Frankly all companies are fucked if some of those events happen so this is nothing to be disheartened about, but you should be aware of them.
Finally, you mentioned doing this at a quantitative shop, what is your average holding period for these instruments? Do you account for transaction costs? or are you looking at end of day prices on specific days only. Would you have a stop loss? At any point during the past 10 years did any of these instruments break the stop loss, thereby forcing your hand to sell and incur a loss at that point?
I don't want to be negative, but any arsehole can tell you yea that'll work quit your job and do it, that's easy.
Good luck anyway T
appreciate the constructive criticism...not to shut you up but just to clarify/give a little more light
on a quarterly basis - outperformed 28 out of 39 quarters on avg by 825bps. underperformed 11 out of 39 quarters on avg by 525bps
looking at large cap companies...theoretically could scale to very large
tends to underperformed in down markets and outperform in up markets underperformed by 1700bps relative in 2008 but outperformed by 5600bps relative in 2009 toying with a long/short that outperformed in 2008 but didn't outperform by as much in 2009
average holding period is a little more than 1 month. model does not account for transaction costs but shouldn't be that significant of an impact since the turnover isn't ridiculous (i.e. short term positions like minutes long). looking at end of day prices on specific days only. no stop loss
thanks for your thoughts
Just by this, I wouldn't want to invest. Anyone can do well in an up market and even beating the index as well. BTW looking through crystal ball to see if it is going to a up and down market is just like random walk.
It looks like you are in some way leveraged against the index, i.e. you are long in the larger components of it, so upswings and downswings overly impact you.
Quant shops don't trade 1 month durations, unless its instantly hedged and something special happens at the end of it.
Thats the negative side. The positives, if you can reverse the thing and do the same for negatives you may have something. So when it goes down you make money and when it goes up you lose, a bit, then run them at the same time and you have a strategy viable for an electronic trading company.
If you can make this thorough and meticulous etc. it shows you all the things you need to show, apart from the idea not working. Save the coding and comment it and subsection it thoroughly then you can copy and paste it into the next one so you can build that superfast.
sd? drawdown? sharpe? s&p plunged by 40% in 2008, if you underperformed by 1700bps, you scored -57%?
If you underperform in downmarkets you're not going to have a viable strategy except for very patient, trapped capital. most institutional investors are generally not happy with underperformance in markets where everything else is down, since they will look to HF strategies to provide diversification.
Effectively what you've described sounds like just adding leverage to the index (in terms of returns; not in terms of what you do since I have no idea what that is). What's your overall correlation to S&P?
Also, keep in mind that investors are much more wary of quants these days, so if you have the problems that you're mentioning it's going to be especially tough to raise funds
If I understand the return profile correctly, your best bet would be a quantitative mutual fund, since a non-absolute return strategy is not going to appeal to HF investors throwing down 2 and 20
For better or worse, institutional investors care about risk management and diversification more than alpha. Show them you have low volatility, low complexity, and low correlation across all asset classes and they might get excited.
Illum impedit in et laborum. Reiciendis quo molestias nihil eum odio totam architecto. Cum soluta pariatur nemo sed eius qui. Totam facilis ut eaque illo magnam qui eius.
Ratione cumque maxime impedit. Est aperiam veritatis sed quod.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...