At first sight, the results show that you have split pairs - basically what should be a pair on pins 1&2 is using one wire each from two different pairs. Ditto for the 3&6 pair. Effectively what it means is that both ends are mis-wired - probably the white/green wire in pin2 and the orange wire in pin3. This is something that the cheap "flashing light" testers will not show you.
The near end crosstalk test (NEXT) shows bad crosstalk between the two pairs according to the graph you have posted but it's not shown in the results table (there should be 6 lines for all the combinations of pairs). The fact that it's not as bad as I'd have expected does lead me to think that perhaps there's some very bad wiring practice been done rather than genuinely split pairs. it may be that someone has (as I've seen done) untwisted loads of wire length - the twists (both within the pairs and between pairs) should be maintained right up to the terminations.
NEXT is a measure of "when you are shouting down one pair, how much of that can you hear in the pair where you should just be hearing the other end" and is possibly more critical than the simple attenuation your test results show.
The TDX graph shows a lot of "mess" in the first (I assume) 100ft of cable. Perhaps this has been badly handled (kinks and the like) which is causing some of the problems. It's not been stapled in place has it ?
TBH, it's a bit confusing to me. NEXT is bad, but not as bad as a genuine split pair problem would produce.
Pairs 4&5, and 7&8 will show massive losses because they aren't connected at all - in principle there should be zero signal reaching the other end, but crosstalk in the connecting leads and test gear at each end will give you a small amount.
But, I'm puzzled by the lack of some figures in the results - e.g. "No Reflection" instead of a length. What length is the cable ? Was the right equipment used at the far end ?
