How accurate are the Leaderboard performance stats?
TradeKing has worked hard with our partner Maxit to make Leaderboard stats both an accurate reflection of any single investor’s performance over a given time period, as well as providing a solid basis for comparing two TradeKing clients’ returns. At the same time, calculating performance stats does involve some art (along with a lot of mathematics). The following is a brief explanation of the assumptions used in calculating Leaderboard stats for TradeKing clients.
Money-weighted versus time-weighted
Performance stats can be calculated based on either a money-weighted or time-weighted approach. Money-weighted calculations factor in the timing and size of cash inflows and outflows, while time-weighted methods do not. Time-weighted methods are frequently used by professional money managers, who don’t control cash inflows or outflows – those are controlled by investors themselves. However, since individual investors do control when they add or subtract funds to an investment, money-weighted calculations arguably provides a truer picture of an individual investor’s true performance.
TradeKing uses two money-weighted calculations, Modified Dietz and Internal Rate of Return, to generate the performance numbers for each investor’s Leaderboard figures.
If you added funds to your TK account just before an upswing in your investments, the money-weighted approach will factor in that decision, resulting in a higher performance result. On the other hand, withdrawing funds before that upswing will result in a penalty to your performance results.
Similarly, adding funds before a downturn in your portfolio value is also weighted negatively – although if you’re ”buying on the dip“ to lower your average cost basis, your longer-term performance stats will eventually reflect any rises in the investments over time.
Correcting cash skews
Sometimes you may see TradeKing clients with performance stats for both a long- and a short-term timeframe – say, 6mos and 5-days – but with ”N/A“ displayed for the timeframes in-between. What’s up?
This occurs whenever the calculation is greater than 1,000%. It’s intended to cull results that are wildly skewing due to large cash flows in or out of the account during the period in question. Of course, it’s also theoretically possible, if unlikely, that the client transferred no cash but simply had more than a 1,000% gain or loss during that timeframe.
Trades outside of a TradeKing brokerage account
The most important caveat to keep in mind about the TradeKing Leaderboard is this: the Leaderboard only reflects performance in a client’s TradeKing account; it does not include performance in other brokerage accounts.
Does that mean it’s possible to stash all your losing trades in an outside brokerage account and migrate all your winners to your TradeKing account? Not exactly. If you transferred losing trades out, the TradeKing Leaderboard would capture that value as a sale and lock it in as negative performance results. When winning positions are transferred in, they’re treated as a cash inflow, effectively making it impossible to transferring a winning record in, so to speak.
I’m ready for the math. How specifically are Leaderboard stats calculated?
TradeKing uses a money-weighted calculation, Internal Rate of Return (IRR), to generate the performance numbers for each investor’s Leaderboard figures. There are a number of methods to calculate a money-weighted return, and IRR is widely considered to be one of the most accurate.
Time-weighted methods are frequently used to measure the performance of professional money managers, who don’t control cash inflows or outflows – those are controlled by investors themselves. However, since individual investors do control when they add or subtract funds to an investment, money-weighted calculations like Modified Dietz and IRR arguably provides a truer picture of an individual investor’s performance.
Here’s an example of how IRR is calculated:
On April 1, you invest $10,000 in Stock XYZ.
On May 1, your investment is still worth exactly $10,000, and you decide to increase your investment by $5,000. You now have a total investment of $15,000 in XYZ.
On June 1, your investment has doubled in value from the previous month and is now worth $30,000.
The initial result is calculated using the Modified Dietz method:
(EMV – BMV – CF/BMV + Adj. NCF) * 100
Next, the Internal Rate of Return (IRR) is calculated as follows:
BMV - EMV/(1 + r) + ΣCFi/(1 + r) w ≤ ± 1 basis point
Where …
BMV = Market Value Beginning: Market Value + Accrued Interest
EMV = Market Value Ending : Market Value + Accrued Interest
Adj. NCF: Adjusted Net Cash Flow
w = day#/period
CF = Net Cash Flows
r = Return
In money-weighted calculations, the timing and size of the $5,000 deposit does impact the results. If you had deposited your $5,000 earlier in the timeframe discussed, your performance figure would drop. The longer the money sits in the account without increasing its value, the less the impact of the cash flow.
How money-weighted calculations compare with time-weighted figures
Using the same example from above, a time-weighted calculation would factor out the impact of the $5,000 cash inflow. As you review the calculations below, keep in mind a ”subperiod“ is created every time a cash flow in or out occurs. Those subperiods are then geometrically linked, or combined in to generate a value over the entire period.
Subperiod calculation:
Ending Market Value = Beg. Mkt. Val. + Net Cash Flows
Geometric Linking calculation:
[(1 + Subperiod Return1) x (1 + Subperiod Return2) x … (1 + Subperiod ReturnN) -1] x 100 = TWR
Generating the following Subperiod Results:
TWR for April 1st – May 1st : 0%
TWR for May 1st – June 1st: 100%
The two subperiods should then be geometrically linked to generate the final result:
TWR for Apr 1st – June 1st: 100%
A time-weighted calculation would factor out the impact of the $5,000 cash inflow, yielding a lower performance result.
Want more?
We created the TradeKing Leaderboard to promote transparency and accountability between investors sharing ideas through our Trader Network. In a phrase, transparency is credibility. That's why we welcome the opportunity to explain our Leaderboard's calculations in more detail to any client who'd like to get under the hood a bit.
Modified Dietz
[(30,000 – 10,000 – 5,000)/(10,000 + 2500)] * 100 = 120%
IRR = 125%
IRR is a more complex, iterative calculation which can’t be shown here. Its purpose is to refine and improve the Modified Dietz result.
-1 x 100
Subperiod I
(($10,000/ (10,000 + 0)) -1 ) x 100 = 0%
Subperiod II
(($30,000/ (10,000 + 5,000)) -1) x 100 = 100%
Geometric Linking
[(1+0) x (1 + 1) -1] x 100 = 100%