any guide here to Charles' MFO Ratings?

linter

Have there been posts about this before?
I'm a little confused about how best to interpret Charles' great MFO Fund Ratings. See below for example.
It seems to me you can't use any of the statistics to compare one fund to another unless the funds in question have been around for roughly the same amount of time. In the example below, that'd mean you can compare DREGX, VEIEX, and GBFAX, all with ten years of life; DRESX, THDAX, FRNMX, and AMDWX, with three years of life; and WAFMX and SIGIX, with one year or less of life.
I mean, in the most basic sense, WAFMX has an APR of 22.9, which is great but doesn't mean much given its short life. If I'm right, all you can say is its relative APR is better than SIGIX's by a long shot, but that's the only fund-to-fund comparison you can make.
And if that's true, then the same holds true for all the others, especially the all important Sortino and Martin ratios.
Then again, I could be dead wrong about this.
Thoughts?
(One other small thing. In the footnotes, maybe there's a way to state for which metrics a bigger number is better ... or is bigger better for all of them?)

Charles

Hi Linter. Here's link to Definitions page. Also, if you double-click the table, it should enlarge to full screen. You're absolutely right about the period comparison, so utilize accordingly. The Return Group, however, is directly comparable within same Category. So, in the table above, for example, DREGX out performed VEIEX, GBFAX, and AMDWX for the past three years. Hope this helps.

linter

thanks for that.
(i take it you mean amdwx, not wafmx, right?)

Charles

Reply to @linter: Right!

Derf

Reply to @Charles: I was wondering how often you plan on updating stats, qter, semi, or annual ?
Thanks for all your hard work !!
Derf

Charles

We thought we'd start with quarterly updates, see how it goes. So, just completed the ratings for 7600 or so funds through month ending September, resulting in about 480 Great Owls...or just over 6%.

linter

One thing I noticed is that funds are sorted by Sortino metric grouped by evaluation period. Would it be fair to assume, then, that you consider Sortino to be the most valuable statistic out of the current grouping?
I also wonder if there's not a way to enlarge the group of comparable funds. In the tables I've seen, most funds listed have 5 and 3 year returns but, as it stands, 3s can't be compared to 5s. But what about if you gave the 3yr metrics for the 5-year funds (or at least the Sortino and Martin numbers). That way, the comparable universe would be greatly broadened.

Charles

Reply to @linter: Ha! You sound like Investor. The return rankings are comparable since they are over same evaluation periods, as applicable. They are actually based on Martin not Sortino. The other tabulated numerical metrics (STDEV, Sharpe, etc.), like you note, are only for the longest evaluation period applicable. The issue is practical table size, since there are 7600 funds being evaluated across (up to) 5 different evaluation periods. Maybe it would be practical to print-out these metrics for all evaluations periods on the Fund Risk/Return Profile pages we are working to make available on-line. Hopefully, we will see how folks use them and refine/improve accordingly, like your suggestion here.

linter

okay, things are getting a little clearer in my wee pea brain. but let me ask this: for comparison purposes, wouldn't it be more useful in the return group rankings to give the actual % number, as opposed to the 1 to 5 group number, since the spread, as is, is a large 20 percentage points? probably i am once again missing something obvious. i look forward to the explanation!