Hi all,
I'm doing a bit of research in to multi-system testing.
Just wondering how people allocate equity across systems? Do you aportion the equity for each system? Do you not bother and just allocate equity on a trade by trade basis? What proportion? What ratios? What percentages?......
Thanks in advance.
red
Equity allocation for multi system trading
-
redbullpeter
- Roundtable Fellow
- Posts: 82
- Joined: Mon Apr 28, 2003 9:12 am
- Location: London, UK
Redbullpeter,
Basically, we use a form of risk "budgeting". We start by allocating among sectors. We try to keep it simple. We just try to ensure that it is evenly distributed across sectors.
The percentage of the risk budget allotted to any one system for any given market is based on the combined analysis of the inter-market evaluation model and the sentiment model.
Now isn't that simple?
Sabrinian
Basically, we use a form of risk "budgeting". We start by allocating among sectors. We try to keep it simple. We just try to ensure that it is evenly distributed across sectors.
The percentage of the risk budget allotted to any one system for any given market is based on the combined analysis of the inter-market evaluation model and the sentiment model.
Now isn't that simple?
Sabrinian
Last edited by Sabrinian on Fri Jul 02, 2004 10:06 am, edited 4 times in total.
-
redbullpeter
- Roundtable Fellow
- Posts: 82
- Joined: Mon Apr 28, 2003 9:12 am
- Location: London, UK
-
steady_jake
- Full Member
- Posts: 22
- Joined: Fri Jun 27, 2003 8:05 pm
- Location: Warsaw, Poland
- Contact:
LOL. A great tag line. It takes a really high IQ to come up with some screwy justifications. History shows some very strange ideas amongst the intelligencia (sp??) of nations that the "common" people wouldnt share (takes free time as well 
).
Its hard to see them in real time especially if you share them. Some would point to the IMFs prescriptions for solving economic problems (which were amazingly popular in the 80s and 90s) as a recent example.
John
).Its hard to see them in real time especially if you share them. Some would point to the IMFs prescriptions for solving economic problems (which were amazingly popular in the 80s and 90s) as a recent example.
John
Steady_jake
However, the actual spreadsheets used to do the ranking aren't really that complex. Basically, you have a bunch of risk caps that ensure that you are diversified across sectors and do not have too much open risk in any one sector, market, or system. The reasoning behind this is that there doesn't seem too much diversification to be had within a sector. For example, US financials (US, TY, FV) or currencies (JY, SF, EC) all seem to correlate when you really want the diversification!
Given these risk caps, you have to decide how much to allocate among the various systems. I do not switch one system on or off. I just vary the % risked in each trade according two factors: Intermarket strength/weakness and sentiment. If the risk caps haven't been violated, the % risked becomes a factor of how strong the market is relative to others and whether it is showing signs of being overbought or oversold.
Simpler?
MODERATOR's NOTE: Unchecked the "Disable BBCode in this post" box. Not really sure why that even exists? If you are having trouble with a BBCode like the
Touché! Ouch! You are quite right! But it wasn't a slip. I was aware of this when I posted. It does indeed seem complex when you write out.A Freudian slip, possibly?
However, the actual spreadsheets used to do the ranking aren't really that complex. Basically, you have a bunch of risk caps that ensure that you are diversified across sectors and do not have too much open risk in any one sector, market, or system. The reasoning behind this is that there doesn't seem too much diversification to be had within a sector. For example, US financials (US, TY, FV) or currencies (JY, SF, EC) all seem to correlate when you really want the diversification!
Given these risk caps, you have to decide how much to allocate among the various systems. I do not switch one system on or off. I just vary the % risked in each trade according two factors: Intermarket strength/weakness and sentiment. If the risk caps haven't been violated, the % risked becomes a factor of how strong the market is relative to others and whether it is showing signs of being overbought or oversold.
Simpler?
MODERATOR's NOTE: Unchecked the "Disable BBCode in this post" box. Not really sure why that even exists? If you are having trouble with a BBCode like the
above make sure the BBCode is not disabled by this checkbox. There is also a check in the Profile that sets the default for each user.
Last edited by Sabrinian on Thu Jul 17, 2003 9:22 pm, edited 1 time in total.
-
steady_jake
- Full Member
- Posts: 22
- Joined: Fri Jun 27, 2003 8:05 pm
- Location: Warsaw, Poland
- Contact:
Basically, you have a bunch of risk caps that ensure that you are diversified across sectors and do not have too much open risk in any one sector, market, or system. The reasoning behind this is that there doesn't seem too much diversification to be had within a sector. For example, US financials (US, TY, FV) or currencies (JY, SF, EC) all seem to correlate when you really want the diversification!
Given these risk caps, you have to decide how much to allocate among the various systems? I do not switch one system on or off. I just vary the % risked in each trade according two factors: Intermarket strength/weakness and sentiment. If the risk caps haven't been violated, the % risked becomes a factor of how strong the market is relative to others and whether it is showing signs of being overbought or oversold.
Simpler?
A lot. Now I have read it without "showing signs of being" exhausted.
But seriously, I would not rely on such vague factors like "relative strength" of the market or its being "overbought/oversold" in my risk management models, even if I could define them quantitatively. Of course it is all very subjective, but I would rather keep my assumptions about the nature of the markets separate from what I know works on the basis of statistics and math. The former class would apply exclusively to entries, stops end exits, the latter to money management and risk control.
Only an opinion, though. Or a question of liking.
Take care,
Jakub
Jakub,
I understood that both Relative Strength and OB/OS are determined by mathematical means.
Sometimes it is good to feel exhausted from learning. Easily consumed sound bites are often junk food.
I think Sabrinian's post was quite a contribution and introduces some integrated methods to determine asset allocation. And for the empiricists, I have seen the test results and the curve is smooth.
I understood that both Relative Strength and OB/OS are determined by mathematical means.
Sometimes it is good to feel exhausted from learning. Easily consumed sound bites are often junk food.
I think Sabrinian's post was quite a contribution and introduces some integrated methods to determine asset allocation. And for the empiricists, I have seen the test results and the curve is smooth.
-
steady_jake
- Full Member
- Posts: 22
- Joined: Fri Jun 27, 2003 8:05 pm
- Location: Warsaw, Poland
- Contact:
Dear Damian, Dear Sabrinian,
I apologize for any sarcasm and thank you for bringing me to dimensions.
Anyhow, I would take this opportunity to try to make the point, I was trying to make, clearer, if possible.
Attributable to my ignorance, I have never seen a working money management model, which would apportion money to trades on the basis of market specific factors like Relative Strength or OB/OS (undisputably being determined by mathematical means) or others for that matter. The only instance of it whas the original Turtles method of relating the position size to market's volatility. I have always had problems with it, though, because my learning about money management was perhaps to heavily influenced by its gambling roots, where there is no place for market specific factors connected in any way with forecasting, just strictly probabilities. Perhaps this is this ignorance that made me reject the model of Sabrinian's a priori, or with just a perfunctory dilligence. I you now of any other "hybrid" models like Sabrinian's, preferably fully disclosed ones, which I might test myself or check how they compare to simpler models based solely on the anti-martingale bet strategy, please do me a favor and post them here.
I was just trying to make a point (I agree that I might have been more polite or less sarcastic - in fact, this sarcasm might have distracted you from my point) that I want to make clear distinction between the system portion of my trading and allocation portion of my trading. I want my expectancy to come exclusively from the trading system design and I want to use simple, robust methods, not influenced by the market considerations in any way (other than particular market's correlation), for deciding the bet size. If something went wrong with my trading, it would be easier to distinguish where does the problem come from.
I hope that I have been able to clarify my point to some extent.
Once again, please accept my appology for dismissing the validity of Sabrinian's contribution in such an uncivilized manner.
Jakub
I apologize for any sarcasm and thank you for bringing me to dimensions.
Anyhow, I would take this opportunity to try to make the point, I was trying to make, clearer, if possible.
Attributable to my ignorance, I have never seen a working money management model, which would apportion money to trades on the basis of market specific factors like Relative Strength or OB/OS (undisputably being determined by mathematical means) or others for that matter. The only instance of it whas the original Turtles method of relating the position size to market's volatility. I have always had problems with it, though, because my learning about money management was perhaps to heavily influenced by its gambling roots, where there is no place for market specific factors connected in any way with forecasting, just strictly probabilities. Perhaps this is this ignorance that made me reject the model of Sabrinian's a priori, or with just a perfunctory dilligence. I you now of any other "hybrid" models like Sabrinian's, preferably fully disclosed ones, which I might test myself or check how they compare to simpler models based solely on the anti-martingale bet strategy, please do me a favor and post them here.
I was just trying to make a point (I agree that I might have been more polite or less sarcastic - in fact, this sarcasm might have distracted you from my point) that I want to make clear distinction between the system portion of my trading and allocation portion of my trading. I want my expectancy to come exclusively from the trading system design and I want to use simple, robust methods, not influenced by the market considerations in any way (other than particular market's correlation), for deciding the bet size. If something went wrong with my trading, it would be easier to distinguish where does the problem come from.
I hope that I have been able to clarify my point to some extent.
Once again, please accept my appology for dismissing the validity of Sabrinian's contribution in such an uncivilized manner.
Jakub
-
steady_jake
- Full Member
- Posts: 22
- Joined: Fri Jun 27, 2003 8:05 pm
- Location: Warsaw, Poland
- Contact:
Dear Damian, Dear Sabrinian,
I apologize for any sarcasm and thank you for bringing me to dimensions.
Anyhow, I would take this opportunity to try to make my point clearer, if possible.
Attributable to my ignorance, I have never seen a working money management model, which would apportion money to trades on the basis of market specific factors like Relative Strength or OB/OS (undisputably being determined by mathematical means) or others for that matter. The only instance of it whas the original Turtles method of relating the position size to market's volatility. I have always had problems with it, though, because my learning about money management was perhaps to heavily influenced by its gambling roots, where there is no place for market specific factors connected in any way with forecasting, just strictly probabilities. Perhaps this is this ignorance that made me reject the model of Sabrinian's a priori, or with just a perfunctory dilligence. I you now of any other "hybrid" models like Sabrinian's, preferably fully disclosed ones, which I might test myself or check how they compare to simpler models based solely on the anti-martingale bet strategy, please do me a favor and post them here.
I was just trying to make a point (I agree that I might have been more polite or less sarcastic - in fact, this sarcasm might have distracted you from my point) that I want to make clear distinction between the system portion of my trading and allocation portion of my trading. I want my expectancy to come exclusively from the trading system and I want to use simple, robust methods, not influenced by the market considerations in any way (other than particular market's correlation), for deciding the bet size. If something went wrong with my trading, it would be easier to distinguish where does the problem come from.
I want such a money management or risk control model that would work whether my assuptions about the state of the market, it's future direction and so on turn out to be true or not.I bet you can understand that.
Hopefully, I have been able to clarify my point to some extent.
Once again, please accept my appology for dismissing the validity of Sabrinian's contribution in such an uncivilized manner and without necessary scrutiny. I hope it will spark up even more discussion and will bring good to everybody in the end.
Jakub
I apologize for any sarcasm and thank you for bringing me to dimensions.
Anyhow, I would take this opportunity to try to make my point clearer, if possible.
Attributable to my ignorance, I have never seen a working money management model, which would apportion money to trades on the basis of market specific factors like Relative Strength or OB/OS (undisputably being determined by mathematical means) or others for that matter. The only instance of it whas the original Turtles method of relating the position size to market's volatility. I have always had problems with it, though, because my learning about money management was perhaps to heavily influenced by its gambling roots, where there is no place for market specific factors connected in any way with forecasting, just strictly probabilities. Perhaps this is this ignorance that made me reject the model of Sabrinian's a priori, or with just a perfunctory dilligence. I you now of any other "hybrid" models like Sabrinian's, preferably fully disclosed ones, which I might test myself or check how they compare to simpler models based solely on the anti-martingale bet strategy, please do me a favor and post them here.
I was just trying to make a point (I agree that I might have been more polite or less sarcastic - in fact, this sarcasm might have distracted you from my point) that I want to make clear distinction between the system portion of my trading and allocation portion of my trading. I want my expectancy to come exclusively from the trading system and I want to use simple, robust methods, not influenced by the market considerations in any way (other than particular market's correlation), for deciding the bet size. If something went wrong with my trading, it would be easier to distinguish where does the problem come from.
I want such a money management or risk control model that would work whether my assuptions about the state of the market, it's future direction and so on turn out to be true or not.I bet you can understand that.
Hopefully, I have been able to clarify my point to some extent.
Once again, please accept my appology for dismissing the validity of Sabrinian's contribution in such an uncivilized manner and without necessary scrutiny. I hope it will spark up even more discussion and will bring good to everybody in the end.
Jakub