sp_ cmon you can also do a fair crowdfund every time you have double digit improvements rather than selling 5% improvements for 0.1 BTC
I could also wait 6months and publish nothing. My work is mostly opensource. The numbers are real. If you compare my work to the old version of ccminer you will see that my kernals are 50-100% faster. I have done this in 1.5 year.. I have spendt 1000's of hours..
Without waiting 6 months you could at least wait to have something you were able to test before releasing it (which is clearly not the case in half of your releases... cf the revert you apply to the other half)
The problem is that you are keeping comparing to that old version. However, things have evolved a lot since then (not only on your side).
If some of your kernel did get faster that's because me and other have developped new techniques (or recover some stuff from Christian which had been overlooked and for that I am thinking to keccak... since you were talking about it...).
I never denied that you did a great job on x11 and quark but that was year ago, since then what you have been doing, mostly, has just been waiting for other to release stuff, fiddling with it, and exploit it to your own benefit.