Although the major averages have continued to rise over the last three weeks, the homebuilder sector, previously a highflier, has been getting crushed. Most of the builder stocks peaked on or around Feb. 13, and since then their decline has run the gamut from orderly to ferocious.

As is typical in a correction, stocks that rise the most during the ascension tend to drop back the most when the trend changes. For example, Ryland Homes (NYSE:RYL), which rose an incredible 114% from its low in October 2011 through mid-February, has pulled back 13% in the last three weeks.

It’s also common for the highest-risk, lowest-priced stocks in a group to be the most volatile. Thus, Beazer Homes (NYSE:BZH), which blasted up 184% from a low of 1.38, to a high of 3.93, has pulled back 18% in the same timeframe. That may seem like a large decline, but it represents only about 10% of the stock’s furious bull move — far less than the usual Fibonacci retracement percentages, such as 23%, 38%, 50% or 66% we would expect after such a steep advance.

On the other hand, Toll Brothers (NYSE:TOL), which was up 74% for the October-to-February bull run, has endured a relatively small and orderly 4.5% pullback thus far.

Comparing these stocks with the group average, the PDR Series Trust SPDR Homebuilders Index (NYSE:XHB) rose 62% during the same period and has corrected a mere 2.6%.

So when this correction ends, which stocks should investors buy?

Let’s look at a table of some of the most widely held homebuilder stocks. I have divided the companies into two groups: conservative (higher-priced issues with smaller gains and smaller declines) and speculative (lower-priced issues with larger gains and larger declines.

In the conservative camp are stalwarts such as Toll Brothers and Lennar Corp. (NYSE:LEN), along with XHB.

In the speculative group are the more volatile stocks, such as Beazer, Hovanian Enterprises (NYSE:HOV) and Pulte Group (NYSE:PHM).

A few conclusions can be drawn: First, while the percentage decline is nearly identical for both TOL and LEN, LEN’s gain was nearly 12% better than TOL’s from the low to the peak. Therefore, LEN is the stronger of those two.

Furthermore, the difference between TOL and XHB in both percentage gained and declined is small enough that it makes owning TOL a more risky investment. Instead, one could own the entire XHB index and benefit from the reduced risk of diversification without giving up substantial gains.

On the other hand, the difference in percentage gain between LEN and XHB was substantial enough, and the percentage of decline small enough, to make owning LEN worthwhile. Therefore, if we are choosing the more conservative approach to owning homebuilder stocks, we want to own either LEN or perhaps the XHB index.

Second, among the more volatile stocks of the group, HOV is clearly superior to BZH at the moment since it has a greater percentage of gain and a smaller percentage of decline since October.

Comparing HOV to PHM is almost as simple. Although HOV has declined 4.2% more than PHM over the last three weeks, during the four-month period of gain, it beat PHM by almost 30%! I’ll take that reward-to-risk ratio any day.

So the stocks to buy once this correction has ended are either LEN or the XHB for a more conservative approach, and HOV as a more speculative play.

But are we near the end of this correction yet? And if the sell-off still has more to go, how much longer should investors wait before they can safely enter long positions? A look at the charts should answer those questions.

Chart courtesy of stockcharts.com

As I’ve indicated previously, LEN could easily retrace back to its rising 50-day moving average at 21.90. But if it fails to hold at the 50-day, there isn’t much support below that until $19.00 (red arrow).

For HOV (see chart below), a Golden Cross, where the 50-day moving average crossed above the 200-day M.A., occurred about a week before the peak. However, more negatively, there is a decline in both the RSI and Stochastic that appears to be headed lower. Furthermore, there has been fairly substantial volume on the down days recently.

The next reasonable support target would be about 8% lower, at the 50-day moving average of 2.38. That would put HOV’s correction at about 20%, a more typical pullback number. If the support at the 50-day M.A. doesn’t hold, then a further decline, to around $2.00, is likely.

Chart courtesy of stockcharts.com

So to summarize, more decline among the homebuilder stocks seems likely.

First, the sector has fallen off a cliff, while the general market has still been rising. Second, the correction has so far been only a small retracement of the four-month run-up. And third, the RSI and Stochastic Oscillators have not yet bottomed out and turned upward.

Given that the overall market seems overbought and perhaps ready for a correction, I would not want to bet that one of the weakest sectors in recent weeks would suddenly turn around and head higher while the rest of the market begins to fall.

So let’s wait a bit longer before we initiate any long positions in the homebuilder stocks. But when we do, either LEN or HOV would seem to be the two best stocks to play.