Tally for March '06

Sort by newest Sort by date
I think that time to expected retirement (TER), which depends on many personal decisions and values, is the most likely correlate to highly risky moves. In general, typical asset allocations with longer TER have a greater amount of investements in stocks, whereas those with small TER have a greater amount invested into bonds and money markets. The model has been adopted by TSP L funds. In general, the longer the TER, the greater amount of historical return and volatility.

If you plot the TER for TSP members, I suspect that the correlation will hold given a sizeable sample size. Thus, long TER --> greater returns, probably greater allocation into S and I, greater historical return.

FYI: 5-year TSP service - 25 year TER - currently 100% I
but power account due to SEP+IRA earlier in my career + maximizing to max IRA every year.
 
It's about the percentage, not the sum. What we are testing is who has the best strategy or luck, not who has the most in their fund. Ultimately, the one who produces the best percentage gains over time will likely have the largest amount of cash as well (over time).

Besides, we can all tell who has the best percentage gain strategy by the posted allocations. To determine who has the most cash, wouldn't one have to submit quarterly earnings statements?

As far as weighted risk, etc., wouldn't one also have to assess an individual's risk tolerance and factor that in.... too much work IMO.

It's not broke, don't fix it.
 
I agree it's really useful to track "who has the best strategy or luck" - and the longer the period of tracking, and the more consistent the tracked individual is about their strategy, the more meaningful tracking based on who has the best percentage gains.

However, it seems to me the value of MDD or other attempts to quantify risk undertaken is that it allows more meaningful comparison of different individuals with shorter track records. And it helps control for strategies/styles that might work great in certain types of market trends but not others - even if you only have a short tracking period. At least that's why I think the professional money managers seem to use something more than just percentage gained/lost for their assessments.

Now do I believe we should further complicate the tracker - No! Just sharing my view that there is real value in a more layered analysis of performance - some may undertake this on their own....

Oldschool
 
oldschool said:
However, it seems to me the value of MDD or other attempts to quantify risk undertaken is that it allows more meaningful comparison of different individuals with shorter track records.
Click to expand...

The problem is there is no real way to measure risk. Whichever method you choose it's based on what happened. Risk has everything to do with what might have happened. I personally feel it's a little risky being in stocks right now. If stocks go up from here without a significant pull back, that doesn't necessarily mean it was not risky to be in them.

Another 2 cents.

Dave
<><
 
Oops, I'll have to disagree for someone holding a passive portfolio. However, you might be absolutely correct for a market timer.:cheesy:

Most financial researchers equate volatility with risk and assume a normal distribution of returns. Consequently, using historical data, two standard deviations from the mean give you the possible range of returns with 95% probability.

In my case, if the future looks anything like the past in terms of average returns and volatility, I expect to receive an average 11.7% portfolio return, +- 19.4%, with a 95% probability. In other words, I can expect a portfolio return in the range of 31.1% to -7.7%. If that's too risky, I can dial up the bond allocation, e.g. to 50%, and receive a lower average return with lower volatility. If I feel lucky (which I don't), I could dial down the bond allocation to zero, go 100% stocks and receive a higher average return with higher volatility.

For example, with 100% C Fund, you can expect an average 12.6% return, +-17.2% (based on 1970-2005 data). In other words, a return in the range of +47% to -21.8%. We haven't seen that +47% recently, but we saw -22.1% in 2002.

Therefore, using statistics and historical market data, I can predict what my returns will be, not perfectly, but within a reasonable range of values. I can also determine, before the fact, whether or not my portfolio is risky compared to other possible portfolios.
 
Wheels said:
Big If!

Dave
<><
Click to expand...
Perhaps. However, the market data from 1926-2005 is pretty consistent. In addition, I've seen domestic market data back to 1800s that basically tells the same tale. Finally, the purpose of probability theory is to predict the future. Otherwise, we're totally in the dark. :blink:
 
rokid said:
Perhaps. However, the market data from 1926-2005 is pretty consistent. In addition, I've seen domestic market data back to 1800s that basically tells the same tale. Finally, the purpose of probability theory is to predict the future. Otherwise, we're totally in the dark. :blink:
Click to expand...

You may be right. I'm not really thinking of it mathmatically as much as practically. If Evil Knevil jumps a bunch of buses successfully 9 times in a row, and then wipes out bad on the 10th try, does that mean that the 10th try was the only risky one? Was there any less risk for the 9 previous attempts?

As far as investing you said it yourself when you asked, How do market timers gauge risk prior to the outcome?

In any event, I'm game for whatever our trackers decide. I just thought that highest returns was the easiest way to go.

Dave
<><
 
Like Rokid, I have also looked at the data for the last 60-80 years and also think it's pretty consistent. And from what I've seen, a robust, well-tested strategy with historical low risk/drawdowns is likely to have low drawdowns in the future. That's why I think it's important to measure downside risk (not just standard deviation) if you want a better idea of future risk. (Note to Rokid: check out the Sortino Ratio if you want to see a measure of downside risk used in academic finance:
http://www.investorwords.com/5793/Sortino_ratio.html)

Case in point: I just tested a system first published in 1995. One of its key selling points was its low drawdowns. Before its release, it had the following backtested performance from 1942 to 1994:

Compound annual return 10.06% vs. 7.33% for the S&P 500 (dividends not included)
Max Drawdown: 12.94% vs. 48.20% for the S&P 500 (in 1974)

Ok, so what happened if you began trading that system real-time starting 1/1/1995 until today's close? You'd have the following performance:

Compound annual return 11.91% vs. 9.70% for the S&P 500 (dividends not included)
Max Drawdown: 10.80% vs. 49.15% for the S&P 500 (in 2002)

It sidestepped the 2000-2002 bear market in almost the exact same way it avoided the 1973-74 meltdown.

A key tenet of market timers/technical analysts is that markets are the products of human emotions--specifically fear and greed. Since human nature doesn't change, neither do the patterns or relationships that create risk and reward. If a market timer can detect such patterns (and prove it with a significantly long history of low-drawdown results), then my money is on them continuing to avoid excessive risk in the future.

ST








rokid said:
Perhaps. However, the market data from 1926-2005 is pretty consistent. In addition, I've seen domestic market data back to 1800s that basically tells the same tale. Finally, the purpose of probability theory is to predict the future. Otherwise, we're totally in the dark. :blink:
Click to expand...
 
SystemTrader,

Makes sense. If you back test your strategy against historical market data, you can predict average portfolio return and risk. If you don't, you have to learn from experience whether or not you're pursuing a risky strategy.

I'll also check out the Sortino Ratio. Thanks.
 
Am I still being tracked at 30% C, 30% S, 40% I?

Just wondering as I've been there since 10/05.

Again, it would be good to place this up as an aggresive buy/hold allocation vs. what everyone else is doing.

Thanks

Aslan
 
Hi Aslan. How have you been?

I believe "The Tracker" (mlk_man) has required everyone, even the buy and holders, to check in at least once per month into their account so we know you are still around. This is to avoid tracking the people who pop in once then never come back.
 
Gotcha,

I'm fine. I didn't feel the need to post since I was a buy/hold person right now. I've been at that allocation for 6 months, and was just wondering how it stacked up against all the active traders. That's all. No big deal
 
You must log in or register to reply here.

Similar threads

FundSurfer
Tally for April 06
2
Replies
38
Views
6K
Rod
Rod
FundSurfer
Tally February 06
Replies
16
Views
5K
mlk_man
mlk_man
FundSurfer
January 06 Tally
2
Replies
34
Views
7K
rokid
R
Back
Top