But only if you have a time unit from beginning of the project, that would work...as shown in FA@h graph, some batches go quicker, some less quick...if you take 30 day average, you end up with problems!

That's why I suggested on many occasions to techs/admins to use start date of the project...& subtract the days that project has been paused...to use that for a project 6m before ECD statement in Research page...

It has been only said that it's a hard to make such a thing...

I do know we had hick ups with OET...but my ECD are for a long runs better than anything else we have here...if only some people would join in the math to their graphs, things would be so much better looking...

I do keep track of a 7 day average, but it jumps around too much for my taste. I like the 30 day average precisely because it does smooth out the peaks and valleys. I also experimented with some longer time frames, but settled on the 30 day figure because it seemed to track more closely to a trend line. If the project had exhibited greater volatility in such factors as length of work unit or number of crunchers it would be less likely to follow a trend line, especially in the short run. So far it does look like batch 126500 will be issued either very late on October 4 or very early on October 5, so at least for the last month or so this method seems to come close to the right time. I appreciate your input.
Cheers
Edit: Looks like it will be October 5 for 126500.
Edit 2: MCM1_ 0126500_ 6370_ 0-- In Progress 10/5/16 06:01:06 10/12/16 06:01:06

@Sgt.Joe

My comments were directed at ECD calculations in general and not at any specific project. Your process seems to work quite well, so I would stick with it. I suspect a lot of it is personal preference and one's desired accuracy at any one point in time. BTW, do appreciate the work you've put into the MCM progress reports.

The following figures are based on a total number of work units of 530,000,000 from cubes (project scientist) with the addition of Seippel (WCG tech).
Last 30 day average results:507,111
Percentage Complete:???????
The average length of a WU is currently running about 4.08 hours, slightly shorter than last week, and still trending steady around the 4 hour mark.
Still have not seen any of the batches from 122278 to 122402. The ovarian units started with 122403 as the lowest number.My latest batch number was 126610 which would give about 163 batches for the week. Therefore, by this measure the project is about ??.??% complete.
We are still at about ??? more months (approx. ? days)of additional crunching at the current recent run rate.The projected completion date is now a ?. Production decreased a bit this past week going from about 3.517 million WU's to 3.338 million WU's completed. During this past week the time devoted to MCM1 was just over an average of 224 years per day, compared to 234 years per day the previous week. Remember, the situation can change at any time and these are only estimates.
I have left some question marks because there is a lack of information to even speculate.Perhaps one of the techs will be kind enough to chime in with the estimated number of new batches of work. I decided I will not go into a much more speculative projection of the end of project just based on the amount of work done, the percentage listed on the research page and current pace of units being completed. Just too speculative. I wish I could do better, but the lack of information does provide a bit of hurdle.
Only one chart this week, as the second one would be meaningless.

Cheers

+1
I come to this thread for analysis not available on the Research page. I appreciate the work you put in here and look forward to having some solid numbers again soon!