STATS MAKE ME CRY

Hello,
call me an idiot but I can't force post hoc tests make sense in Statistica 8.0. Even when I input "dummy" table containing 3 vars going like "1,2,1,2.../100,200,100,200.../1000,2000,1000,2000" and run simple one-way ANOVA with default setting, what I get when running LSD, Tukey or whatever is a significant difference between vars 1 and 3 but no difference between 1 and 2. Or do I perhaps misunderstand the output (see at the bottom)?
Pls help!

Homare

LSD test; variable Var5 (Spreadsheet in Workbook1)
Error: Between MS = 93519,, df = 27,000
Var6 {1} {2} {3}
1,5000 150,00 500,0
1 0,287153 0,000000
2 0,287153 0,000000
3 0,0000 0,000000

In the meantime, I tried SPSS with similar result - obviously I make a systematic mistake:
If I input the series: 1,1,1,2,2,2/10,10,10,20,20,20/100,100,100,200,200,200 into repeated measures ANOVA, post hoc tests recognize a difference between series 1 and 3, 2 and 3, but not between 1 and 2. The outcome is the same even for Bonferroni's correction, which I suppose works exactly like a t-test with alpha level 0.05/n. And well, performing the analysis with t-test yields (naturally) a highly significant outcome (p = 0.001).
What am I missing?

Homare