speed up query time

[markororke]

i am looking for a quicker way of performing this query:

Select count (distinct t1.child)) from WineHigh t1 join WinHigh t2 where t1.Seed = 14 and t1.depth =3 and t2.depth<3 and t1.child = t2.child and t1.Seed = t2.Seed;

this is used primarily to return the no. of distinct child of t1 that where previously there in previous levels.....

in other words the query is comparing results at level 3 to previous levels to return the number of times they are the same.