Minor Trivial-Draw problem

[tvt2020]

This should be a winnable game not draw.

[Event "test"]
[Site "tvt2020"]
[Date "2011.06.07"]
[Round "-"]
[White "tvt2020"]
[Black "EngineTN2 (UCCI2WB)"]
[Result "1/2-1/2"]
[TimeControl "40/600"]
[Variant "xiangqi"]
[FEN "2ea1keh1/2P6/5a3/2p3p1H/9/9/9/9/9/2E1K1E2 w 0 1"]
[SetUp "1"]

{--------------
. . e a . k e h .
. . P . . . . . .
. . . . . a . . .
. . p . . . p . H
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . E . K . E . .
white to play
--------------}
1. Hh8+ Kf8 {+0.03/6 0.2} 2. cxc9 g5 {+0.35/31 25} 3. Ei2 Ade8
{-9.70/29 18} 4. Ea2 Ad9 {-4.74/26 46} 5. cxd9 c5 {+10000000.00/27 57} 6.
de9 c4 {+10000000.00/25 52} 7. Exc4 Ae8 {+10000000.00/24 11} 8. Ke1 Af7
{+10000000.00/21 1.4} 9. ef9 Ae8 {+10000000.00/19 1.1} 10. fxg9
{Xboard adjudication: Trivial draw} 1/2-1/2

[H.G.Muller]

OK, I see. The black Horse is trapped.

The 'trivial draw' adjudication is only an approximation; it is not supposed to be 100% reliable. (e.g. in Chess it also considers KQKQ a trivial draw, while there are a few quite long wins in that end-game.) If you want to be absolutely sure, you should switch it off, and let the engines play to the end. Or judge final positions of games thus adjudicated afterwards, and if they are not really draw, let the engines finish them later.

In this case WinBoard triggers on considering KH being sufficient defense against KHP. Which in general is true. The given position looks like a study to me, something you would never encounter in a real game.