Have you read the documentation at the Broad's web site (http://www.broadinstitute.org/gsea/doc/GSEAUserGuideFrame.html) or their original PNAS paper?

Quoting from there:

"When you run the gene set enrichment analysis, the GSEA software automatically normalizes the enrichment scores for variation in gene set size, as described in GSEA Statistics. Nevertheless, the normalization is not very accurate for extremely small or extremely large gene sets. For example, for gene sets with fewer than 10 genes, just 2 or 3 genes can generate significant results. Therefore, by default, GSEA ignores gene sets that contain fewer than 25 genes or more than 500 genes; defaults that are appropriate for datasets with 10,000 to 20,000 features. To change these default values, use the Max Size and Min Size parameters on the Run GSEA Page; however, keep in mind the possibility of inflated scorings for very small gene sets and inaccurate normalization for large ones."

Have you read the documentation at the Broad's web site (http://www.broadinstitute.org/gsea/doc/GSEAUserGuideFrame.html) or their original PNAS paper?

Quoting from there:

"When you run the gene set enrichment analysis, the GSEA software automatically normalizes the enrichment scores for variation in gene set size, as described in GSEA Statistics. Nevertheless, the normalization is not very accurate for extremely small or extremely large gene sets. For example, for gene sets with fewer than 10 genes, just 2 or 3 genes can generate significant results. Therefore, by default, GSEA ignores gene sets that contain fewer than 25 genes or more than 500 genes; defaults that are appropriate for datasets with 10,000 to 20,000 features. To change these default values, use the Max Size and Min Size parameters on the Run GSEA Page; however, keep in mind the possibility of inflated scorings for very small gene sets and inaccurate normalization for large ones."

Thanks, Michael and Thomas, I had not seen it. I think it solves my question perfectly.

Yes you are correct

If you still have to use a small gene set (< 25 genes), you can do so, but make sure to validate the results by checking the ES graph. If you are getting a clear trend there, you be seeing something real.