Shannon's Journal Discussion Thread, Vol. 5

[Catman]

I couldn't do that even if I wanted to.  Ben hasn't linked it up yet because I wasn't finished.  I'll have him link it regardless, but it's going to take some time either way.  In the mean time, I'll keep trying, but at this point I suspect something or someone is trying to prevent me from releasing this, and I'm sick of the crap.

These things happen. I believe it's a good thing, here's why...

I think I know what's wrong:

Since it's such a new program, such new technology, so far beyond AOL's level, the models can't even fully quantify what they're dealing with. Like asking a fish to compare a tricycle to a Ducati. So getting a reliable number isn't feasible. That's my view. I've had similar issues in business at times. Sounds like your models need to be built out a bit more to reflect the new potency. This is a very good sign of progress I believe. I bet that's the problem, Shannon. Try to relax, it's more indication of a large step forward here!

I actually think the problem is slightly different.  I think it boils down to the following things:

1. Based on the models giving me answers that don't qualify as "unstable", but also don't qualify as "stable" - which I have never seen before in the 15+ years I have been using them - I suspect there is some sort of interference from my subconscious or someone or something else attempting to inject "noise" into and interrupt the answers to this sequence of questioning to delay or prevent the release.
2. This generation of tech is so far from what we had before that I have no idea what parameters are reasonable to even use to run through the models to find the best usage patterns.  I've had to do my best to make educated guesses on that.  I'm sure if the answers I got were wrong, we will figure it out through adjusting as necessary.

I'm just really tired of and frustrated with a lot of things right now that are slowing me down, holding me back, delaying progress, killing success, etc. and which are either out of my hands or happening for stupid reasons.

Sounds like we're on the same wavelength there with #2. Said differently, but seems the same idea.

So real-world testing will have to provide new modeling data for this gen, 5.75.7G at a minimum, in order to have a new "best practises" config for modeling going forward. Maybe that info will provide a platform going forward that is accurate for modeling, or maybe new info will have to be added at each new step. Who knows, I suppose it depends on just how big the steps remaining end up being. The bigger the steps individually are, the more of a chance for modeling to become unstable again. Which makes sense as it goes out of bounds and can't fill in gaps to make a prediction past a certain point.

Like I said, this is good news regardless. It means the programs have outgrown the old config for the models. Fantastic. That's evolution. We'll straighten out your models with new real-world data with this gen to fill in the gaps and get them accurate again. No problem. It's progress, it's very good news.

Don't stress, this is a positive! Ben will link us up in time, for now, you sit back and play with the dogs and gorge on your steak, lol . We'll get that new data for you.

I think this program is going to be unusual in it's required usage patterns relative to other programs of this generation because it deals with fear, which generally requires more aggressive approach than otherwise.  I was initially expecting a config like 1 loop per day on, 1 day on and 4-5-6 days off per ASRB2 cycle.  The answer I got was considerably more aggressive than that.  But for other programs in this gen that are not going to meet with such a stiff amount of argument, it may in fact be that we can now run 1 loop per week and be fine.  We shall see.  That difference isn't really helping me figure out the parameters though.

Of course, it'd be similar to a DMSI situation due to possible resistance to the goal. That makes sense for sure. To be frank, many times I've seen such ASRB2s, they haven't really moved the needle with people. I've always been hugely skeptical of such ASRB2s. So I'm not really surprised there. Seems those kinds of ASRB2s have been overly conservative. For this gen and beyond though, it seems your models need fresh data to input in order to have a degree of accuracy going forward. We'll get it for you as I said, no worries.

Too many unknowns at this point about future ASRB2s etc. We don't know what we're dealing with yet, although the revelation about the models intrigues me and makes me optimistic for the future. Seems that OF V3 has outgrown the models as I said, that's remarkable. We'll figure this out as we go and get you the new data you need to make informed decisions and model again.

Now all we need is upside-down Ben to come back from his walkabout, find his mate he's waiting for, put on some pants since he's in his jocks, clean up his mate's mom's shop that was wrecked...and fix our links, BRAH!