There is perhaps an over-emphasis on burn speed here. While it is the subject of a great deal of advice and speculation, any good quality scientific study of the issue seems to have been buried in the discussion, hidden as proprietary data by the disk or writer manufacturer's, or was never done in the first place. It does make good sense that a recordable CD with sufficiently sharp and deep pits (actually disturbance of the dye since recordable CD's do not have pressed pits buried in the medium) will have a better chance of playing on a wider variety of disk readers. That is, however an issue of burn quality, not burn speed per se. It is also an issue of player quality. There are older dedicated CD music players out there that will not play a recordable CD from any source, and tweaking burn speed will not make them do so.
The speed hypothesis is based on the idea, that a burn point on a rapidly moving disc is exposed to the laser for a lesser time, and that the power cycle of the laser has a significant ramp up/down time. The first should produce a less deep pit (intense disturbance), and the latter would result in a less distinct burn region. The loss of distinctness would be due to the rising edge of the laser power being spread over a larger linear path, producing a dye distubance density that is more like a sine wave than the desired square wave. If that were the only factor, there would seem to be no need for data.
Unforunately there are other factors that need to be considered, and the relative contribution of those factors to the problem turns a simple solution into an enigma. For one thing, the power cycle of the laser, and the total delivered power are factors. Burners designed for higher speeds have more powerful and faster responding lasers than the ones used in the older burners. Without knowing a great deal more about these devices than the manufacturer reveals, it is difficult to predict the extent to which underburn and indistinct problems will result from a particular speed. There is also the potential for over-burn if a too-powerful laser rests on a point for too long causing spread from the desired point into the surrounding medium--a puddle instead of a pit. In extreme cases this could cross the linear boundary into adjacent "grooves". A similar problem used to affect the cutting of grooves in masters for vinyl records. The "mastering engineer" was the guy with the experience and judgement to be able to control the amplitude of the signal (loudness at various frequencies) driving the device to prevent the cutting edge from gouging into the next track and ruining the master.
The sensitivity of the dye is perhaps the major factor. Blank CD's designed for high speed burning, should be optimized to respond to high speed burning. The dye may well be the most significant factor in determining if a less responsive CD player chokes. Certain dye "colors" have a bad reputation with dedicated CD players. It is not necessarily true that a CD that will not play in every reader has more "errors" than one that will. The reading laser may be respnding to frequency and reflectance as much as to the depth or distinctness of the burn.
This is not just a "quality" argument. A "high quality" CD from a respected manufacturer, may still be optimized (intentionally or not) to work at a specific speed, responding to a specific laser frequency, and burn duration. It is not immediatly obvious that burning such a CD on an older burner at a lower speed will produce fewer errors. The control firmware for the burner may also be a factor as it nominally controls the burn timing. So different burners using different lasters, firmware and media would be expected to produce different results. In some cases the slowest available speed may at least theoretically not be the best.
Ideally, if you are doing production burning, you would test a variety of media and speeds for your equipment using software that can compare the read result to the intended written data. You would also test an impractically large number of players to see if your ideal media and speed combination approaced a universal compatibility. As time goes on, and the CD players are designed to accept a wider variety of media (and older players hit the landfill) this will probably be less of an issue. In the mean time, if you want to keep your customers happy, a liberal return policy might be the best solution.