Originally Posted by ZX-Tex
The only problem I had with MLV and VE analyzer was that it did not seem to want to converge on a solution. That is, each subsequent log and VE tune would go leaner-richer-leaner-richer, and just oscillate back and forth.
In other words, the tune was doing the exact same thing that the mixture does on a stock car normally, yes?
(waiting for it to settle)
After raising the ease from normal, to hard, to very hard, this problem seems to have been minimized.
And that's exactly why that variable is there.
Ever painted a car? Or painted anything that requires sanding?
When you first start out, you have the MS set with large EGO authority and short step times, and you have MLV's analyzer set to easy. This ensures that you can quickly get the table roughed in. That's the 500 grit stage- it's making big changes quickly.
As you progress, you start to slow the EGO feedback, and set MLV to Normal. 1000 grit- starting to take the rough edges off the map. You'll probably start learning to properly hand-tune the cells around idle at this point.
As the map starts to really fall into place, you reduce EGO authority and switch MLV to Hard. 1500 grit- now the map is getting smooth, you'll be filling in those cells at the edges of the map that never get touched.
And lastly, you'll spend some time at MLV Very Hard, just to see what it does- cells will be going up a point or two one day, down a point or two the next. At this point, you can whip out the 2000 grit and give it a nice hand-polish. (ie: tweak the last few out-of-bounds cells by hand).