That poll has it 47% to 43% Clinton when the actual results were 48.2% to 46.1% Clinton. The only real discrepancy is they overestimated the vote totals for 3rd party candidates. The numbers they had when excluding 3rd parties was 49% to 46% Clinton. Do some blending of those two and they basically nailed it. Is this really the poll you want to call out as fundamentally flawed?