According to Public Policy Polling, their polling predicted the recall of Angela Giron in Colorado Senate District 3. They then chose not to release the results.
Whether we believe them or not, this doesn't speak well of the firm.
Consider either scenario. On one hand, we could take them at their word. They polled a race during an off time, that was drawing national interest and big bucks. They got a surprising result contrary to the wishes of the Democrats they tend to serve. So, they chose not to release numbers that could motivate the recall forces and damage the incumbents.
So by PPP's own admission, what critics like me have been saying about them has been true all along: the polls they release, or don't, are cherry picked to favor Democrats, much as individual voters are cherry picked to make their releases. Their name would possibly be better changed to Public Policy Trolling.
On the other hand, they could simply be full of it in this case. Claiming to have had a poll that predicted a surprising result, never releasing it, but then attempting to claim credit after the fact, is a pretty cheap move that anyone could make. Science is about predicting the unexpected, not confirming what we already knew.
Either way, this release by PPP can't possibly improve one's opinion of the firm.
Update: It's one of those rare moments when Nate Silver and I seem to agree on a point, as he and PPP have had a big old Internet fight.