Your "effective fee" is very misleading. [...] So if you really mean what you say and that you should use a comparable metric, advertising a generic "effective fee" is not it.
As with any metric, both nominal fee and effective fee are open to misinterpretation. We don't claim we have the holy grail of metrics. Calling our effective fee 'very misleading' is quite an accusation, as we try to be as transparent about it as possible.
The fee for a pool is one of its key characteristics. Miners rely on that number, so it should be as comparable as possible. Many non-PPS pools silently withhold payment for invalid blocks and are still listing their fee as 0%. Due to the nature of PPS, it is impossible to withhold payment for invalid blocks without calculating it as a 'fee'. The amount per share simply isn't the theoretical (50/difficulty) BTC.
Therefore a 1.5% nominal fee PPS pool can be expected to pay as much as a 0% nominal fee non-PPS pool (ceteris paribus). That's not immediately intuitive, especially to people new to the mining scene. If a newbie sees a PPLNS pool at 0% nominal and a PPS pool at 1.5% nominal, he thinks the latter one costs more, while in fact they cost the same.
This proves that the nominal fee, while clearly defined, is open to misinterpretation. The notion of effective fee that we use is not very clearly defined, and that's a disadvantage which we try to mitigate by explaining what
we mean by it everywhere we list it. While less clearly defined, in terms of USD costs it is a better fit than the nominal fee. Perfect? Certainly not.
On our site and in our initial forum post we list both the effective-
and the nominal fee, and an explanation. From those pieces of information everyone should be able to make up their mind about whether it's a good deal for them. You may not agree with our interpretation of effective fee, but at the moment I don't know how we can be more transparent about it.
[...]
If you want to advertise a comparable metric, use that. At XX Hashrate, the fee is YY USD per month.
We would encourage anyone to build a tool that takes as input one's hashpower and outputs a histogram of expected payouts, in BTC or USD. Such a tool would combine all reward-influencing factors into one graph: nominal fee, the variance each type of pool has, and the effect of stale rate and incentives like invalid block compensation and stale share compensation. Maybe we'll even build it ourselves, because I'm sure we'd come out looking pretty good.
If you'd like to talk about some objective reward-efficiency metric that all pools can use to advertise themselves, please start a new forum thread for that.