Quote from: sturle on February 15, 2016, 07:55:19 AM

Quote from: kano on February 15, 2016, 02:14:43 AM

No, it's just way less likely, that's the point it seems you've missed.
That's the whole point of using 95% vs 75%
Poisson distribution statistics aren't linear ...

It less less likely to be triggered by a minority (<50%), but it is in no way failsafe against getting triggered by less than 95%, as Organofcorti points out.

Relevance zero - no one but a fool would say what "WILL" happen.
Ooc's post is all about probabilities and confidence, not what "WILL" happen as you have YET AGAIN made the same mistake below.
I've also posted to him what he missed regarding the fact that nodes convert before they produce a block.

Quote from: sturle on February 15, 2016, 07:55:19 AM

Quote from: sturle on February 14, 2016, 09:01:40 PM

Quote from: kano on February 14, 2016, 06:34:52 PM

Edit: see point 4.

See my previous posts. I wrote a summary of point 4 above:

Quote from: kano on February 15, 2016, 02:14:43 AM

Quote from: sturle on February 14, 2016, 09:01:40 PM

Quote from: sturle on February 14, 2016, 12:15:55 PM

I made no argument against 95%, btw. However 95% of the last 1000 blocks will be triggered by less than 95% (see explanation by Organofcorti). 95% of one difficulty period (2016 blocks, fixed measurement points) is much more likely to be a certain 95%. At 95% of the last 1000 this somewhat premature activation doesn't carry a high risk of a prolonged fork, so it probably doesn't matter as long as there is a cut-off date where the soft fork will fail if not achieved.

No, you've made the clear mistake of taking probabilities and saying something "WILL" happen.
Obviously something can happen, but if you decide a low probability is an issue, then you can never do a fork, since you can never reach 100% certainly that every miner has switched.
You choose a high confidence interval.
75% is low.

Yes, it will happen if the hashrate is stable on both sides. If the divide is 94.5% vs 5.5%, and the 5.5% share have bad luck for a day, it will trigger. Since the test is done every ten minutes by average, the chance of the 5.5% having bad luck for long enough at some point is very high.

Sigh - OK you've lost the plot.
Nothing is "IT WILL HAPPEN" - even you've just said "if .... something else happens" so that's not even a "IT WILL HAPPEN"