Once upon a time... wait, wrong story format. I was playing around with some write strategy changes. I have a spindle of RITEKR01 discs rated for 2.4x that I have yet to really successfully burn at 8x. I was also playing around recently with improving my RICOHJPNR01 burn quality by having my burner use the RICOHJPNR02 write strategy when burning RICOHJPNR01 discs. The results of this swap of the Ricoh strategies was fantastic, so I decided to do the same for my RITEKR01 strategy--I made it so that my burner will burn RITEKR01 discs with the RITEKR03 strategy. And the results were wonderful! Low PI and PO. I happily popped my newly 8x-burnt RITEKR01 into my combo drive to do a reading scan. And much to my shock, not only did the drive start thrashing at the end, I got an unrecoverable read error! Highly unusual, as my successful RITEKR01 burns (at 6x) read smoothly on my combo. What went wrong?...
Since every RITEKR01 that I've ever burned at 8x have coastered, despite low PI/PO, I started to think that there must be something else at play in dictating readability. What about ADER?
So I did something fun; I cleared out my EEPROM's power calibration data. (Please don't try this at home :p) And I burned a RICOHJPNR01 disc with this really messed up EEPROM. Needless to say, I got extremely high PI/PO errors, and the disc was a bona fide coaster, but interestingly, I got very low ADER errors when I later scanned the disc using the ADER test.
Interestingly, the RITEKR01 also had very low ADER error rates when I scanned it... except at the end, near the place where the unrecoverable error occurred. There, the ADER rate jumped.
Although this has already been said in the ADER/ATER thread, there is more that goes into quality than just PI/PO BLER tests. You have two factors at play here: the power of the laser and whether or not things are being written at the right place...
Coaster #1: RITEKR01 - This disc coastered because of a jump in the ADER rate. Causes of an ADER-related problem: poor precision of the laser's tracking, resulting either from poor hardware (in which case, ADER rates will be high for all discs) or from discs that are physically unsuitable for high speed burning (the grooves are not cut well enough, there's more warping towards the edges, etc.)... however, the write strategy was good, the power calibration was good, and that's why PI/PO remained low for the most part.
Coaster #2: RICOHJPNR01 - This disc coastered because of (deliberate) improper power calibration. This could be caused by a number of factors: poor writing strategy (since the strat controls the amount of laser power used), poor drive calibration (if the factory didn't quite get the EEPROM calibration right when the drive was shipped... or if you purposefully nuke your EEPROM, as I did), or the disc's dye has problems (perhaps the write strat was designed for one type of dye, and then the manufacturer suddenly decides to switch to a different dye type without changing the media code). In this case, bits were being written at the right place (because the RICOHJPNR01 discs have better grooves or whatever and was simply better suited physically for higher speeds), but the power being applied was wrong, which ultimately coasters the disc.
And another thing...
Something to add to all this is that if you have good media that burns with really high cliff-like PI/PO errors, that's probably indicative of power calibration problems, which means that a firmware update (with a newer, better write strategy) could solve your problem.
Alternatively, if you have a coastered disc, and you run an ADER scan on it, and it turns up with very low numbers, then you know that the disc coastered because of improper power calibration (which can be more easily fixed, as a new firmware could do the trick) and not because of poor hardware precision (which could prove to be harder to rectify)
And now, for some pictures...