Suggestion: maximum drawdown

Questions about FantasyTSP or suggestions for improvements should be directed here.

Moderator: Aitrus

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Suggestion: maximum drawdown

Post by Peaceful-B »

I think maximum drawdown (MDD) is an important performance measure. Though I am not sure how to best integrate it into the ranking system on this site. One idea is to have several MDD bands and rank users by return within each band. Or, sort users from lowest to highest MDD, and show both their MDD and their return. Or something else. How to best include MDD is something to think about, but I do think it's a good idea to include it.
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

User avatar
TSPking
Site Admin
Posts: 1066
Joined: Thu Nov 15, 2007 7:02 am

Post by TSPking »

Please explain further as to what MDD is or provide a link with explanation so I can see how it is calculated.

Thank you for the suggestion.
TSPking

It's a gift...and a curse ~ Adrian Monk

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

MDD is the highest percentage drop in equity. Higher MDD is indicative of a riskier strategy.

Let p(t) be the price at time t.

DD(t,s) = -min( 0, p(t) / p(s) - 1 ), where t > s (this is important).

MDD is the maximum of DD(t,s) over all t and s (again, with t > s).

When calculating MDD, in theory, you could consider all t and s such that t > s. But note that, for making calculations quicker, you only need to consider s that are local price maximums. In other words, you only need to consider s such that p(s) >= max( p(s-1), p(s+1) ). Likewise, you only need to consider t that are local price minimums. p(t) <= min( p(t-1), p(t+1) ), t > s.
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

Also, of course, you need to consider s = 1 and t = the latest time period available.
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

User avatar
TSPking
Site Admin
Posts: 1066
Joined: Thu Nov 15, 2007 7:02 am

Post by TSPking »

So in terms of FantasyTSP, I suppose we could extend the Leader Board to include additional ranking categories that consider evaluation of MDD in addition to returns when determining rank for various time ranges.

The question I have now is what is the relationship between MDD and returns? Would they have equal weighting when determining rank?

Could you also provide calculated MDD for Peaceful-B for the calendar Month October which I can use as a reference when attempting the MDD calculation for FantasyTSP?

Thank you.
TSPking

It's a gift...and a curse ~ Adrian Monk

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

Like I said, I am not sure how to best add MDD to the leader boards.

In my experience, a strategy with an MDD of up to about 35% is considered to be a good general strategy. A more risk-averse good strategy would have an MDD of up to about 15%. These are MDD's based on as much data as possible, not based on 1 month or even 1 year. So one idea is to have rankings of returns filtered by MDD. For example, a ranking of by returns regardless of MDD, a ranking of returns with MDD <= 35%, and a ranking of returns with MDD <= 15%.

Since you probably want rankings on shorter-term time periods, these MDD cutoffs (15%, 35%) would not apply. One idea, and I think it's a good one, is to base MDD cutoffs on the MDD's of the TSP funds themselves. For example, L2040 is sold by the TSP as being a good general fund. So, you could use the MDD of the L2040 fund as a MDD cutoff. This will give you all the funds that are less risky than L2040 (or whatever other fund you pick). Then, the people who have MDD <= the MDD of L2040 and the return > the return of L2040 could be considered to have performed well. They are making more money than L2040 while taking less risk. ... Something along those lines. I'm just brainstorming...

As for my own MDD, I know it's low, but I will have to get back to you. I don't have the files in front of me.
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

McLovin
Posts: 127
Joined: Fri Oct 24, 2008 5:37 am

Would this penalize

Post by McLovin »

people for taking greater risk or reward them. As a business manager I don't mind if my people
take risks just don't lose. On the other hand if the point is to just gain the most then then we can all
sit in I and S (Jumping in and out as the market flows). Kind of like checkers. Boring. Adding MDD
would make the game more like chess... Interesting. Might be something to try. Checkers for some.
Chess for others.

McL
"What we do in life echoes for an eternity" Maximus from Gladiator

"Without knowledge and understanding good decisions are only lucky guesses" McL

User avatar
TSPking
Site Admin
Posts: 1066
Joined: Thu Nov 15, 2007 7:02 am

Re: Would this penalize

Post by TSPking »

McLovin wrote:people for taking greater risk or reward them. As a business manager I don't mind if my people
take risks just don't lose. On the other hand if the point is to just gain the most then then we can all
sit in I and S (Jumping in and out as the market flows). Kind of like checkers. Boring. Adding MDD
would make the game more like chess... Interesting. Might be something to try. Checkers for some.
Chess for others.

McL


I don't think it would penalize because it would not replace the current Leader Board categories based only on returns. Instead, this would be additional Leader Board categories that would calculate rank based on returns and assumed risk (I think).
TSPking

It's a gift...and a curse ~ Adrian Monk

User avatar
Pocono13
Posts: 541
Joined: Mon Dec 22, 2008 4:40 pm

MDD

Post by Pocono13 »

Peaceful Does your model use MDD or something like it to measure risk? So if I understand, a 20 percent gain with <15 percent MDD (risk) would be much better than a 25 percent gain with <35 percent MDD.

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

I think MDD-based leader boards would not penalize anything. They would just give people more choice over which strategy they want to follow. If all they care about is return, then they can use the current leader boards. If they want some risk control, they could use the MDD < 35% leader board. If they're scared by the market or near retirement, they could use the MDD < 15% board.

Pocono13, whether 20% gain with a low MDD is better than a 25% gain with a higher MDD is more or less a personal decision. I think that a good approach to investing is to pick whatever risk level that you are comfortable with, for example, as measured by MDD, and then try to maximize returns.
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

Diesel
Posts: 18
Joined: Fri Jan 09, 2009 3:30 am

Post by Diesel »

I myself think it's a great idea and would give others an idea of what strategies to follow. I follow many of the leaders on the boards here in an attempt to learn the most I can. Being able to sort by MDD would allow me to categorize who in reality, is taking the most risk and compare it to the rewards. Very interesting.
Retired US Army
Current Dept of Army Civilian

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

Here's a much easier calculation than what I posted above. Don't know why I didn't think of it right away.

Let M(t) be the maximum equity for time <= t. Thus,
M(1) = p(1)
M(t) = max( M(t-1), p(t) ), for t > 1

DD(t) = -min(0, p(t) / M(t) - 1)
MDD = max DD(t)
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

McLovin
Posts: 127
Joined: Fri Oct 24, 2008 5:37 am

Please don't take this in a negative way. I know wrong way.

Post by McLovin »

to start the message but what is to prove by doing MDD? We are all cognizant of the fact that the G fund has no
risk and the I fund has the most risk. We are also aware that each fund moving G to F to C to S to I has increasing
levels of risk (going least to highest). It states so in the description of the TSP funds if you read them. What is the point we are trying to prove? Is it to beat the market (roughly 13.5% correct me if I'm off) with the least amount of risk? Is that going to be the point of the MDD boards? If so I would like to open a second account. 1) to try to max out my return and 2) Max my return with minimal risk.

McLovin
"What we do in life echoes for an eternity" Maximus from Gladiator

"Without knowledge and understanding good decisions are only lucky guesses" McL

User avatar
Peaceful-B
Professional TSP Advisor
Posts: 167
Joined: Fri Jan 02, 2009 3:54 pm

Post by Peaceful-B »

You're right in that we know that G has no risk and I has a lot of risk. The interesting question, I think, would be where do two different people's strategies stand?
Disclaimer.Not a registered investment advisor. Does not provide any individualized advice. Past performance is not necessarily indicative of future results.Future accuracy and profitable results cannot be guaranteed.©2009-2010 Peaceful Gains,LLC.

User avatar
TSPking
Site Admin
Posts: 1066
Joined: Thu Nov 15, 2007 7:02 am

Post by TSPking »

I put this on my TODO list. Not sure how fast I'll get it done but I'm shooting for late December or January.
TSPking

It's a gift...and a curse ~ Adrian Monk

Post Reply

Fund Prices2024-03-28

FundPriceDayYTD
G $18.15 0.05% 1.05%
F $19.08 -0.06% -0.74%
C $82.21 0.11% 10.55%
S $82.43 0.30% 6.92%
I $42.57 -0.24% 5.95%
L2065 $16.38 0.02% 8.37%
L2060 $16.39 0.02% 8.38%
L2055 $16.39 0.02% 8.38%
L2050 $32.73 0.01% 6.95%
L2045 $14.91 0.02% 6.58%
L2040 $54.38 0.02% 6.22%
L2035 $14.34 0.02% 5.79%
L2030 $47.67 0.02% 5.38%
L2025 $13.15 0.03% 3.43%
Linc $25.61 0.03% 2.82%

Live Charts

Pending Allocations

Under development. For now, you may view Pending Allocations by going to "fantasy TSP" and selecting "Leaderboard sort" of "Pending Allocations".