Chuck Jaffe at MarketWatch (probably my favorite finance Web site, which is why I link to them so much) wrote a column Friday about rejecting mutual funds that had a great short-term run.

Why? Because the great short-term run often means that the mutual fund is heavily invested in one sector that has had big appreciation recently but is due for a downturn. Get in now and you'll probably miss the upside while feeling the full brunt of the downside.

I have had great success in picking mutual funds. I say this to pump myself up because of my not-so-hot track record in buying and selling individual stocks. Maybe I do so well buying mutual funds because it's so much easier to be dispassionate and I can just look at the numbers to see which funds deserve my money.

But which numbers to look at? In my experience, it's best to look at long-term numbers over short-term numbers, but don't diss the short-term altogether. Here's what I do:

Go to MarketWatch's list of the top 25 mutual fund performers in each of the major fund categories. The simple beauty of these lists is that they offer you the top 25 performers over several different periods: 1 year, 3 years, 5 years, 10 years, 15 years and some even shorter term.

Let's say you're fairly young and you're looking for a growth fund, something with the chance of high returns but also a bit riskier.

Go first to the 1 Year tab on the "Growth" list. Print it out. Go to the 3 Year tab, print it out. Go to the 5 Year tab, print it out. Go to the 10 Year tab, print it out. Go to the 15 Year tab, print it out. Now you've got a lot of paper in front of you.

The first thing you are going to do is compare the 1-Year list to the 3-Year list. Find any mutual funds on the 1-Year list that are also on the 3-Year list and highlight them. Cross out the rest of the mutual funds listed on the 1-Year.

Now compare the 3-Year to the Five-Year, and don't completely forget the 1-Year. Look for the funds on the 5-Year that also made the 3-Year. Highlight those and cross out the rest of the 3-Years (except any 3-Years that had overlapped with the 1-Year, those might be under consideration).

Do the same comparing the 5-Year list to the 10-Year list, and comparing the 10-Year list to the 15-Year list.

What you want to do is find the funds that have the most overlap between these different time periods. I've never found any fund that was a top 25 performer for all the periods listed. More often you'll find overlap in the first, shorter-term periods or the longer-term periods, but not both.

Now, rank the funds that have made multiple lists. If any funds showed up on three lists, rank those first. If you have multiple funds that showed up on three lists, rank the ones that showed up on the long-term lists higher--for example, a fund that showed up on the 5-10-15 lists beats a fund that showed up on the 3-5-10 lists. Likewise with funds that show up on two of the lists. A fund that shows up on the 10 & 15 lists beats a fund that shows up on the 5 & 10 lists.

Make a list of the top 10 funds based on this criteria. Then you have to do a little more investigation.

First, some of these funds may be closed to new investors, especially those that have been successful for 10 or 15 years. So you first have to find out what you can even buy.

Second, check fees & loads. Some of these funds are going to have high management fees, which is OK since they're highly successful funds, but I still steer clear of those with higher fees. In the same way, I steer clear of funds with loads--a fee upfront or at fund selling that eats into your returns. Even with a track record of success, paying fees and loads will eat into your return without a guarantee that these funds will continue to perform as well as in the past.

Then, make your choice and send in your money.

I just did this exact exercise with growth funds and ended up buying into the Value Line Preferred Growth Fund (previously known as the Special Situations fund). It was not the highest-ranked fund under my system, but it was top 25 for both 10 & 15-year growth (at an average 12.92% for 10 years and 12.51% over the last 15 years). The expense ratio is 1.13% which is very low for a managed fund, and there are no loads, so you're not giving away your gains. Plus, you can actually buy into it, which I couldn't do for some of the other potential funds that had closed to new investors.

It takes more time explaining this method than actually doing it. Give it a try and I think you'll be happy with your long-term results.