RHP games explorer -- stat problems?

http://redhotpawn.com/gamesexplorer

In the Master Games section of the database, after 1. e4 c5 2. c3 e6 3. d4 d5 the games explorer shows 50 games featuring 4.e5, and of these White is shown winning 24%, Black winning 34%, and 42% drawn. However, after actually clicking on 4.e5 there are now shown not 50, but 673 games; and it is clear that Black is not the winner of a plurality of them.

Can someone explain the mechanics of the Games Explorer to me? Why does it show that there are 50 games having 4.e5 when offering 4.e5 as a branch point, but 673 games having 4.e5 after the branch is actually selected? And how do I go about finding the 50 games of which Black wins a hefty plurality?

Has the operation of the Games Explorer recently been changed? Now it suddenly reminds me of chessgames.com's Opening Explorer, which was always confusing.

Originally posted by Mark Adkins
http://redhotpawn.com/gamesexplorer

In the Master Games section of the database, after 1. e4 c5 2. c3 e6 3. d4 d5 the games explorer shows 50 games featuring 4.e5, and of these White is shown winning 24%, Black winning 34%, and 42% drawn. However, after actually clicking on 4.e5 there are now shown not 50, but 673 games; and it is clear that Black is ...[text shortened]... ? Now it suddenly reminds me of chessgames.com's Opening Explorer, which was always confusing.

As I understand it the positions are kept in the database by FEN, but you access those positions move by move.
The transpositions of the opening moves may lead to these variations in numbers - I'd guess that the particular sequence you outline arrives at a position that is more commonly arrived at by another sequence of moves.

There are glitches in the current incarnation of explorer as well, I know that certain sequences get mashed because of errors in the FEN - case switching for example, leading to some opening lines that cannot be followed through.

Originally posted by orangutan
As I understand it the positions are kept in the database by FEN, but you access those positions move by move.
The transpositions of the opening moves may lead to these variations in numbers - I'd guess that the particular sequence you outline arrives at a position that is more commonly arrived at by another sequence of moves.

There are glitches in the ...[text shortened]... EN - case switching for example, leading to some opening lines that cannot be followed through.

Well, 4.e5 is a transposition in this case from an infrequent line of the Sicilian into a much more common French (Advance). I would like to see the problem addressed, however. It was quite a rude awakening, though the problem was partly my fault in that I was hasty and didn't follow through sufficiently far in examining the line in question -- always a mistake when using game/move databases.