I did also a forecast, but only based on history:
Development of difficulty since 15.09.13
Dates days to solve 2016 difficulty increased in %
15.09.13- 26.09.13 11 112628549 --
26.09.13-7.10.13 11 148819200 32,13
7.10.13-17.10.13 10 189281249 27,19
17.10.13-26.10.13 9 267731249 41,48
26.10.13-06.11.13 11 390928788 46,02
6.11.13-18.11.13 12 510929738 30,07
18.11.13-30.11.13 12 609482680 19,29
30.11.13-11.12.13 11 707408283 16,07
11.12.13-22.12.13 11 908350862 28,41
22.13.13-03.01.14 12 1180923195 30,01
03.01.14-14.01.14 11 1418481395 20,12
14.01.14- ?? 1789546951 26,15
= average time to solve 2016 Blocks: 121/11 = 11
= average difficulty increase: 316,94 / 11 = 28,81 %
Forecast
dates days for solving 2016 difficulty increased in %
14.01.14-25.01.14 11 1789546951 26,15
25.01.14-05.02.14 11 2305115427 28,81
11 2969219182 28,81
11 3824651228 28,81
11 4926533247 28,81
11 6345867476 28,81
11 8174111896 28,81
11 10529073533 28,81
11 13562499618 28,81
11 17469855758 28,81
11 22502921202 28,81
11 28986012801 28,81
11 37336883088 28,81
17.06.14-28.06.14 11 48093639106 28,81
I was to lazy to put in all the dates :-D
We've spent a lot of time on a Behavioral Model to try to make some sense of what's going on. It pretty soon became clear that there is tipping point where miners will stop buying new kit, because they simply will never make a profit. The difficulty can't keep going up exponentially as a result (unless the Bitcoin value follows the same trend).Our prediction shows difficulty in mid-June at around 18,000,000,000, 25,000,000,000 by end 2014. At that time a TH will earn about $400 a month and eat a good portion of that in electricity costs. As a result, it's worth less than $1.3 per Gigahash/sec.