Blood on the streets...

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Macro
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Sat Dec 03, 2011 11:22 am

Blood on the streets...

Post by Macro » Mon May 14, 2012 1:42 am

=
Best,
Macro
Last edited by Macro on Thu Mar 31, 2016 8:41 pm, edited 1 time in total.

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Mon May 14, 2012 5:19 am

DEpends on if you think human nature has changed - it hasnt
therefore it may just be a case of riding this crap out for another 1 week - 18 months before teh markets let go and unwind in spectacular fashion whereupon i imagine trend following will clean up as it did in 2008 and all the short vol / mean reversion traders will blow up - as they did in 2008
Nothing has changed

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Mon May 14, 2012 8:07 am

I have to admit, Macro's post made my morning. Obviously drawdowns are not fun but to see this type of language (although I cannot really tell who it is coming from) is something that pops up every drawdown. Noone can really predict what will happen but I sure would rather be in my position than in JPM's, who are stuck in their il- liquid large trade that is moving against them.

Macro
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Sat Dec 03, 2011 11:22 am

Post by Macro » Wed May 16, 2012 3:13 am

Chris67, while I do agree that human behavioral biases haven't really changed, I must admit that I'm under the notion that markets may have altered from their steady state of persistence, possibly due to increasing intervention but I may be irrationally pointing fingers here. As a result, I hope to devise a 'mean-reversion' (using the term loosely here) that can complement a trend following strategy in the pursuit of lower volatility and a smoother equity curve.

Toosday, that particular person in question is a portfolio manager at a macro fund and a former MD at an investment bank.

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed May 16, 2012 5:13 am

I have recently added a mean reversion strategy or two - in small size - to the overall portfolio - its bloody tough to find them that offer anything worthy and i ceratinly wouldnt walk away from the mean reversion strategies and just let them run as I would do with trend following systems

I concurr that as of now it certainly looks like the intervention of the authorities is putting a dampner on things - how long will it last ? well i suppose it could go on years ?
Also - and this is me pointing fingers - it truly does appear that market particpants are more stupid than they were - you only have to watch CNBC or watch market reactions to "obvious " pieces of news to work that out
If people think Greece will stay in teh Euro (it may well this time bu it will be gone by Christmas) or that Spain is not in a position that in some ways is worse than Greece and will follow teh exact same fate with consequences 13 times bigger - then they should go see a doctor

its a very tough environment and if you run 250 mill USD and have 25 mill in teh bank im sure its no problem - but when your anew fund 18 months in that simply cannot get its head above water - its a real challenge
C

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Wed May 16, 2012 10:55 am

Chris,

You post is applicable to me as well. It has been a tough time. I can say that it has focused me and caused me to put a significant amount of work into my systems. In the end I am better for it. Sometimes I feel that trying dozens new ideas on the system and having none of them work is worse than drawdowns.

I have been going through my process which helps me cope with the emotions. If you ever want to blow off steam, send me an email.

P.S. Be careful with the mean reversion strategies. We get paid to be tough and there is no such thing as a money machine. Best of luck.

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed May 16, 2012 12:28 pm

i feel sickened by the whole thing at present
it isnt the draw downs that hurt - not to expect them would be ludicrous as a trend follower - its the outlook going forward that really hurts - I cannot see this market environment altering course for a while and that is teh ugly bit - it would be quite easy to see markets just do what theyve been doing for the last 12 months for the next 5 years

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed May 16, 2012 12:29 pm

but then again .. what the hell do i know

(letting off steam)

Macro
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Sat Dec 03, 2011 11:22 am

Post by Macro » Wed May 23, 2012 12:36 am


Jason Mc
Roundtable Knight
Roundtable Knight
Posts: 105
Joined: Wed Dec 31, 1969 7:00 pm
Location: Sydney

Post by Jason Mc » Mon May 28, 2012 9:24 pm

It's re-assuring to know that I'm not the one one finding the current market difficult.

The question is, what to do about it.

I've been trading a medium to long term trend following system since July 2008. Last year was the worst CAGR for my system for the length of my data going back to 1983 (CAGR was down 10% - only the second losing year since 83). It also contained the largest drawdown, down 24%. This adds to the disturbing concentration of large drawdowns since 2006 - four 20% DDs, compared to four 20% DDs for the entire period of 1983 to 2006).

I'm all for the notion that human nature doesn't change. But is the data telling me something is different?

I can improve system performance by moving the ATR entry trigger further way from the current price. While it improves performance over the past couple of years, it decreases performance in earlier times.

So what is better, go with what has worked best in the recent past or stick with the better parameter over the longer term.

Ordinarily, I would go with the latter. However, my concern is the change in operational performance (larger and more frequent drawdowns than ever before) is a sign something is wrong.

Is anyone else experiencing this issue or have some thoughts on the matter?

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Tue May 29, 2012 3:34 am

Jason

I think maybe splitting your cap in half and trading 50% in what has worked best in the Long term and 50% what has worked best in the last 5 years maybe best ?

You have to be careful not to end up chasing performance - to a small degree I did that a bit - i shifted to a much longer term TF strategy in late 11 and wouldnt you know it they got blown out the water in 2012 whereas actually some short term TF stragegies are up money in 2012 - it really is sods law
Have a mix is the best pholiosophy
Also half your capital until something gives

BuyHigh SellLow
Roundtable Fellow
Roundtable Fellow
Posts: 50
Joined: Wed Apr 27, 2011 12:46 pm
Location: U.S.

Post by BuyHigh SellLow » Tue May 29, 2012 12:16 pm

I, for one, have seen a pretty wide set of reactions to LTTF's poor recent performance. Interestingly, many of the CTA's are shifting away from it to at least some degree.

Here's my problem with trend following CTA's shifting into more diversified strategies:

Are they really built to provide alpha outside of their traditional trend following framework? A TF CTA who has been doing TF all along might not have success building and trading an additional suite of systems. If you show me a guy with 20 years experience trading TF systems who has hit a rough patch and is now adding other strategies to the mix, why should I believe this is going to generate alpha? These TF CTA's who are ONLY NOW rebranding themselves as multi-strategy CTA's are entering a lion's den imo. They might think TF is competitive, but I see this multi-strat thing as competing with even tougher groups like Citadel, RenTec, DE Shaw, Duquesne remnants, you name it.

Oh, and say goodbye to that "non-correlated" TF stuff. August 2007 was bad for both quant funds and TF. We'll see even more of this coming. Same for equity-TF correlation.

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Tue May 29, 2012 12:36 pm

Buy,

Your comments on alpha are interesting. IMHO you need to be very precise in defining alpha. Like Sharpe it has become a marketing tool. I think the mixture of people on this board would have a trouble specifically defining alpha. Not from a textbook definition but the differences between their own alpha and their clients alpha. I know the textbook definition but that is not what I am discussing. What I mean is that many traders who do no manage money have a personal alpha much different than money managers. If you are managing a personal portfolio, alpha has a different context than if you are managing others money. The call option of managing others money creates a personal alpha for fund managers much different than for traders managing a personal account. What I believe we are witnessing is the short-term breakdown of client and individual traders alpha versus the money managers who collect management fees. Management, not performance, fees are a great diversify-er against portfolio losses when the losses are coming.

Sorry if I am rambling but as someone who in this life and in past lives has marketed alpha, I find that the math is decidedly in favor of those who collect fees to offset losses when they come. In summary, marketing alpha often results in "de-worsification" (as opposed to diversification) for the client. I think the reason for this is that investors focus on correlation which is an average of a random process and lose focus on the tails.

BuyHigh SellLow
Roundtable Fellow
Roundtable Fellow
Posts: 50
Joined: Wed Apr 27, 2011 12:46 pm
Location: U.S.

Post by BuyHigh SellLow » Tue May 29, 2012 1:50 pm

I think the fund managers are very aware of the call option aspect of what they're asking clients for. Their response, often, if they have a long track record or deep pockets, is to tell you in their marketing materials how much proprietary capital they have. I mean, these are some big numbers, folks. Graham has $1 billion plus, the older TFers have $200 million+, Winton who knows, QIM $400 million. And, you have to think this has a placating effect on those big, nervous institutional investors. And, it certainly hurts the smaller, new managers.

Re:alpha...I think TF alpha is incredibly rare. Research says there's a beta in managed futures, but what's really being said is that there is a beta in trend following.

So, this means what? A simple, mechanical TF system can mimic gross TF CTA returns? Yikes. So, TF "alpha" was just a kind of beta all along. And, now these TF managers are suffering rough beta TF drawdowns and are being forced to try new strategies. And, as investors, we're supposed to buy into this...I sure wouldn't absent some really compelling stuff. Otherwise, how weak is that sales pitch? "Hi, investor. You've been paying me for alpha 2 and 20 while I have just been an expensive way for you to get beta. You'll see on this slide that I am in a maxDD. My new plan is to try to find alpha in ways I have little experience. Hold fast. Sign here. 2 and 20. Thank you."

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Tue May 29, 2012 2:24 pm

I agree that they are aware of the call option. Kovner said as much 20 years ago in Market Wizards.

I think you and I are in the same spot in saying that, unfortunately as an individual investor, you have to find your own alpha. You have to be aware of those promising you a chance at alpha when they are really locking themselves into some alpha. At least with a HF you have a chance at alpha where there are so many "opportunities" for negative alpha right now...ie long term govt bonds and savings accounts. :D

rabidric
Roundtable Knight
Roundtable Knight
Posts: 243
Joined: Mon Jan 09, 2006 7:45 am

Post by rabidric » Tue May 29, 2012 3:36 pm

good points guys.

I wonder if what we are really trying to do with TF, is take a high beta process and filter some of the negative parts of the distribution. I would not feel comfortable calling the result "alpha" though.
"filtered beta" maybe?

Any given filters( i don't mean a MACD) seem to degrade over time no matter what you use, so eventually you probably end up with bog standard beta if you stand still too long with your strategies.

In my experience so far alpha is something that only exists truely where there are barriers to entry. Not many barriers left in finance these days, and any that do arise, fall faster than ever.

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Tue May 29, 2012 4:03 pm

rabidric,

I agree about the filtered Beta. For my entire career, I have tried to work against decisions that focus purely on the mean and consider the entire distribution of prices. I think, based on my understanding of alpha and beta, that these two items focus on the mean. Linear regression (used to determine alpha and beta) is just like a filter, it discounts tail events in favor of more frequent observations. I think the tail observation phenomenon is why TF has worked in the past.

However in this lies the problem, correlation drives many investment decisions. I am sure many of us have heard that diversification is the only free lunch that Wall Street gives you. I think right now the market has taken this away as well. Similar to your comment on barriers being broken. Given that the entire investing world has moved to a risk on/risk off trade (at least based on my TF observations) it seems correlation has trended to 1. I view this almost as I would view volatility in the options market. It will change at some point but since you can directly trade volatility you can profit if vol rises to very large levels and then decays. Maybe people smarter than me can fade correlation as it reduces to its normal levels but I have yet to figure this out. I am trying though.

Correlation in itself does not produce TF returns but it reduces the cost of the call option we all buy. The cost of this option is very high right now and the range traders are collecting the premium.

This is a good discussion and I thank everyone for their thoughts.

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Tue May 29, 2012 8:43 pm

Don't forget that in futures, every long has got a counterparty who is short. Thus it is impossible for everybody to have the same position. Thus it is impossible for the correlation between every pair of trader's positions to equal +1.

By definition, half of the positions are long and the other half are short.

Every "Right Honorable Pure Trendfollower" in Chris67's proprietary Index Of Pure And Honorable Trendfollowers, has got a counterparty. When an RHPT loses money, her counterparty makes money. Obviously. When a huge percentage of all RHPTs lose money, their counterparties make money. Obviously.

So if you think that Right Honorable Pure Trendfollowers are going to lose money in the summer of 2012, figure out what positions they most likely have, and then go the other way. Fade them. Become their counterparties. Somewhere on this site is a correlation study done at NewEdge which reveals an ultra simplistic mechanical system (a 2SMA crossover if I recall correctly) that is correlated with trendfollower indexes at R > +0.85. (R and not Rsquared, again if memory serves). So use this to decide what positions most trendfollowers have - AND FADE THEM.

If you've really got Huevos Grande, fade the trendfollowers selectively. Use discretion, and human judgement, and your own gut instinct to decide which RHTF positions to fade and which not to fade. Also consider using discretion to pick a better entry time and/or exit time. It's well known that trendfollowers enter late and exit late -- perhaps you can improve upon that.

Toosday
Roundtable Fellow
Roundtable Fellow
Posts: 70
Joined: Fri Jun 03, 2011 11:21 am
Location: Austin, TX

Post by Toosday » Tue May 29, 2012 9:38 pm

Sluggo,

Thanks for the response. I am guessing I did a poor job of explaining myself. I am not looking to fade trendfollowers, I am trying to figure out a way to fade the correlation that exists in the markets currently. I do not have a hypothesis that TF will lose money this summer, I actually am hoping for the opposite. For right now, I have cut back trading due to the correlations. If I see something diverge from the current risk-on risk-off paradigm then I may act on it although systemically I have had trouble finding a good decoupling method.

Based on my research, TF has made money in both high correlation and low correlation markets. Unfortunately the risk premium is significantly higher during periods of high correlation, resulting in larger swings. If I cut positions during high correlation periods there is a performance hit and I want to try and avoid prediction.

Sorry for rambling. I appreciate your input. In the end, if the object is to stay in the game then I should not pay the higher risk premium, a subject touched on you historically with your thermal BBBO approach. So maybe I have answered my own question. Thanks again.

Moto moto
Roundtable Knight
Roundtable Knight
Posts: 427
Joined: Mon Jun 01, 2009 4:12 am
Location: once again in the UK

Post by Moto moto » Wed May 30, 2012 7:08 am

topical maybe from the FT FM on Monday.....

image resized by moderator
Attachments
smart_beta_resized.jpg
smart_beta_resized.jpg (304.41 KiB) Viewed 6739 times

Post Reply