Birth Of A New ThumbTack Fund: GFH Fund

I’ve spent the past several weeks since I’m back, trying to figure out the various methodologies, and what is the most commonly used (amongst professional money managers, that is) ways to track returns.

Yes, it sounds trivial, but I’ve been taking it really seriously. I’d explain why later.

As with the bulk of my DD done these days, I’m kinda lazy to type out my thought processes and findings, BUT I’d share my thoughts really briefly.

Basically, I’ve come to realized that the Time Weighted Returns (TWR) and the Money Weighted Returns (MWR) can be drastically different for the same portfolio. Interactive Brokers calculates it automatically for me, and that’s how I realized it.

Check this out:

797) MWR.jpg


798) TWR.jpg

Both these sets of returns are taken at the same time.

My global portfolio shows an annualized MWR of 14.42%, but only an annualized TWR of 10.73%.

So, which one is it? Which one should I really utilize to track returns.

For some reason, my TWR always tends to be lower than the MWR (as far as I’ve observed).

I still haven’t really figured out what that means.

Anyhow, I won’t spend time going into detail of how to calculate each. Most people won’t bother anyway.

The difference is certainly significant, and I’ve previously erroneously thought that the difference wouldn’t quite matter. That it’s most likely in the decimal spots.

But it is significant. I’ve observed, on occasion, a difference of almost 9% once.

I’d share a comment by “kc2024” on this, and a link to a spreadsheet he created to test out TWR.

799) Comment.jpg

Now, for the uninitiated, all this is probably ancient Greek to you. And in all fairness, I think for the retail investor without too large a portfolio, it’s… not very important to figure it out.

But once your portfolio crosses the 7 digits mark, I think it’s actually significant.

I mean, a 10% difference between TWR and MWR would mean a difference in quantum of $100,000 in a single year! (OK, roughly. I know there’s the effect of compounding, and there’s also the effect of the time duration of the capital)

Let me illustrate all this with an example.

Bill has $100k sitting in his bank account at the start of the year, that he can use to invest.

In Feb, he bought 10,000 shares of company ABC at $2, and sold the shares in April at $3. In May, he received a cash windfall and decided to set aside $50k for future investments. In Sept, he bought 10,000 shares of company DEF at $5, and sold the shares in Oct at $6. That same month, he took out $20k from his investment fund to buy a car. In Nov, he bought 10,000 shares of company GHI at $2, and they’re worth $3 as of Dec.

How would you calculate his ROI?

Method 1:

This is what I suspect most retail investors would do:

Total capital invested in ABC, DEF and GHI: $90,000

Total capital from divestments + value of 10,000 shares of GHI as of Dec: $120,000

ROI = (120,000 – 90,000) / 90,000 x 100% = 33.33%

Method 2 (Money Weighted Returns):

800) MWR.jpg

ROI = 23.56%

Method 3 (Time Weighted Returns):

801) TWR.jpg

ROI = (1.25 – 1.0) / 1.0 X 100% = 25%

As we can see…. the ROI figures are very much different.

(OK, this is not a great example. Not sure why MWR and TWR is quite close here, my numbers given in IB are always very far apart)

In any case, the time value of money is mainly what causes the discrepancy between method 1 and the other 2.

We can also see that using TWR is incredibly complex if one has multiple cash injections or withdrawals. Also, it requires the calculation of “units” issued each time there’s a change in your portfolio. And that is really such a big hassle such that it renders TWR mostly impractical.

At this point, most people would be saying…. who cares?!

Well, I do actually, cos I want to know which benchmark to use so that if I’m comparing against an index, or against a money manager, I know it’s a like for like comparison.

Most money managers are judged by TWR. (Yes, the complex one)

Now, the real reason why I’ve been spending all that time trying to figure TWR out… is cos of a new fund that I get to play around with – let’s just call it the GFH Fund. I needed to figure the way returns would be tabulated, as that affects… well, my personal remuneration. Something very important isn’t it?

GFH Fund will be seeded by an investment holding company, owned by yours truly, and some friends. I get full investing autonomy with regard to the fund. Well, frankly, the other folks don’t even care (unless I start losing too much money!)

There’s no real mandate to bother about, but logically, I’d be sticking to what I’ve been doing. No reinventing the wheel here. So it’d be mainly a long/short global equities fund.

And as I type this right now, GFH fund is finally underway.

802) Game on!.jpg

I’ve just seeded it with SGD 100K, with the remaining SGD 400K coming in, probably by the end of the month.

For GFH, these are the stuff that I’ve looked at in the previous couple of months. I wrote this, a few weeks back, don’t think too much has changed:

803) list.jpg

And since I’m really too busy to start posting long investing thesis here, I figured I’d just take the short cut and post my transactions here. Perhaps weekly, or at the min, monthly.

And of course, my dear TWR as calculated by Interactive Brokers.



  1. Hi TTI,

    Perhaps I can give you a scenario and see which method is a better reflection of your true return.

    Ace bought 10,000 shares of ABC at $1 in Jan.
    The share price fell to $0.50 in Jun and Ace injected capital of $5000 (10,000 shares of ABC) to double up.
    All 20,000 shares of ABC were sold in Dec for $0.75.

    Ace break even in this case.

    Method 1: 0%
    XIRR: 0%
    TWR: -25%

    It is easier for funds to plot TWR to show a chart to investors over multiple years, but XIRR will always show an annualised figure regardless of timeframe. This is the main issue why funds use TWR. Anything less than a year, XIRR will show a magnified figure, but it gives a more realistic compounded annual ROI in the long run.

    Of course. if you intend to compare with funds, then TWR will give you an apple to apple comparison.


    1. Hi Matthew
      Yes, I understand your example. Thanks for putting that out here.
      In fact, here’s a much more exaggerated example of what you explained: (Example B)

      Which shows a massive positive TWR despite having a quantum lower than the injected capital.

      As you’ve mentioned, TWR would give a like for like comparison with funds.
      And yea, I guess for the retail investor, IRR aka MWR would be more reflective.



  2. Hi, I came across TWR and MWR when playing with robo platforms like syfe and Moneyowl. Then I found out that the formula they use to calculate TWR is different. I end up tracking their performance all by XIRR to level the ground. I mean…IB TWR may not be the same as what industry TWR is using as TWR seems to have variations. Just my observations here. Thanks for sharing your thoughts all these years.


    1. Hi Lewis

      How can TWR be calculated “differently”?
      There’s only 1 formula to do so, not sure how there can be variations.

      Anyway, IBKR is used by many professional funds, and I’d think their calculation of TWR is legit, since that’s what funds use to gauge their returns vs their benchmarks.

      Personally though, I use MWR.
      I think it’s more reflective for the individual investor.
      TWR is used by funds because they have no control over the funds coming in at any 1 time, and also, TWR allows them to be benchmarked more accurately.
      We can have a situation whereby there are losses, but TWR is actually positive.
      I don’t think that’s what the individual investor managing your own monies would consider as a win… so I think MWR is a more accurate reflection.


      1. Hi, you are right :-) I just reconfirmed the formula from syfe and moneyowl, Syfe wrote as TWR = [(1 + HP^1) x (1 + HP^2) x … x ( 1 + HP^n )] – 1 and I thought it was to the power of n but in fact it is the same as how Moneyowl writes it.

        Too bad thinkorswim doesn’t plot like IBKR. It’s really good to be able to see the performance of our strategies in charts.

        Just for reference:
        Syfe –
        Moneyowl –,being%20influenced%20by%20investor%20cashflows.

        Liked by 1 person

  3. Hi TTI,

    Mind sharing more about the legal structure you used for this fund/compensation?

    Heard that doing a proper fund can cost, even at it’s cheapest, 50k per annum


    1. Hi Dylan
      My personal fund.
      Not a public fund soliciting public monies.
      Yes, there’s a cost to doing so, and I’d think it’s way more than just 50k per annum.
      1stly, you need an MAS license for that. Now, there are ways around that, it involves tagging along the license of others that are approved to manage capital. Not sure if I’m allowed to go into details here. lol

      So yeah, I sure as hell don’t need MAS’s approval to manage my own funds.



      1. Ah okay so basically it’s just a normal company, but with your compensation as a function of the return right?


        1. not sure which 1 u are referring to.
          I have a personal fund, set up with Interactive Brokers. It’s just my personal holdings, that I manage and Interactive Brokers track it. The fund report is posted here and in the “results” tab

          I also have a holding company. It’s an incorporated company, with various shareholders, who’s main function is investment holding. It’s also involved in healthcare operations.
          So yes, if you are referring to that, that is a “normal company”. (My comp is a bit more complex and I don’t really wanna discuss my personal comp here. This is also NOT my full time job btw)



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s