Jorge Stolfi said another stress test had started but there are only a few thousand unconfirmed transactions tonight. It might be the reason for the delay the OP has. I ignored it until now and hoped the test would stop before it did any damage. I hope it doesn't last for days like the last time when there were 10 thousand unconfirmed transactions.
Another "stress test" just started. The backlog is already 60 MB of unconfirmed transactions, and the input traffic is ~ 3 MB every 10 minutes, or ~5 kB/s . So far, the tester is paying 0.2 mBTC/kB, like the last one. Will it keep at that, or increase the fees later?
How accurate are these figures? Its a simulation of a model of the system, so doubly removed!!.
As the dev admits, " the main tradeoff, however, is the dependence on the model assumptions"
AFAIK, the "simulation" is only for the top graph (what fees you would need to get expectation of N minutes to first confirmation), and for the dashed line (capacity) on the bottom graph. The backlog (solid line) and incoming traffic (dotted) on the bottom graph are real data.
The information about the fees being paid by teh unconfirmed transactions comes from
this other site. The big step at 0.20 mBTC/kB must be largely due to the test transactions.
Right now there appeared another large step at ~0.50 mBTC/kB; that may be test transactions, too. However, those transctions are still below the 750 kB level, so they may affect only those transactions (a few thousand?) that paid between 0.20 and 0.35 (beware that the graph is changing all the time). But they may be a sign that the "tester" is raising his annoyance goal...
-snip-