In his article "How to Create a Monster" (Dragon Magazine #276, reprinted at the front of the Monster Manual II, and available online here: http://www.enworld.org/cc/downloads/main.html), Skip "the Sage" Williams presented comprehensive standard statistics for each of the 17 different monster types in the D&D game. For each monster type, a table was given showing normal ranges of abilities, hit dice, and damage, for each possible size category, which many gamers have found to be of great use in designing and converting their own monsters.
However, behind the numbers lurks what may be a significant misstep: the range of Hit Dice for each monster type and size is very much contradicted by the range of statistics given for the variety of monsters in the official Monster Manual. Williams writes that a "creature's total Hit Dice should fall within the listed range unless it is unusually formidable", but by this measurement, practically of the creatures in the Monster Manual are "unusually formidable".
To see this, one should first note that fully 15 out of the 17 monster type tables are given a single, identical Hit Dice progression. This is shown below:
(We can see that in this standard progression, each size category basically doubles the range of the previous one -- except at the anomalous "Huge" category, which multiplies by x4 the "Large" range just before it.)
So, that being the case, what are the only two exceptions to this progression? One, there are Elementals, which double all the hit dice from Fine up to Large (thereby cleaning out the anomalous jump at Huge, establishing a perfect doubling throughout the progression). And two, there are Dragons, who have their unique, insanely powerful increase in Hit Dice compared to other monster types of the same size -- quintupling or more (between x5 and x12) the Hit Dice from the standard progression shown above.
That's all well and good. However, when one compares these standard ranges to the Hit Dice for monsters detailed in the Monster Manual, a disconnect is apparent (see also this page: www.superdan.net/dndmisc/dragons_are_too_small.html).
* Excludes Dragon-types.
Looking at these numbers, we can see that average Hit Dice for creatures in the Monster Manual, from the Tiny to Large size categories, falls outside the standard ranges given in the Williams article -- by a factor of fully x2 over the suggested high-end. The actual maximum Hit Dice for these size categories is anywhere between x5 and x9 over the high-end specified in the Williams article's tables. (The minimums, however, can be said to fit perfectly with those given in the article.)
How did this come to pass? My theory would be based on the observation that the "standard progression" low-end numbers (once a full hit die has been reached) match up precisely with the basis given for "Extra Hit Die" calculation given in the Monster Manual (p. 11). I have to assume that someone looked at those EHD numbers and assumed that most monsters would actually fit within those ranges, when in fact they do not. (This theory is further bolstered by the fact that the first Monster Manual Web Errata released by WOTC inserted an exception to EHD calculation for Elementals, which suddenly alters the EHD base numbers to match the low-end numbers for Elementals' Hit Dice in the Williams article.)
Finally, how should the average DM respond to these observations? I would strongly recommend that anyone creating their own monsters keep an eye to the actual averages from the Monster Manual -- if you faithfully follow the ranges specified in "How to Create a Monster", you will be generating monsters significantly under-powered compared to those in the core rules. In order to simplify this process, I present a suggested revision to the standard hit dice progression:
This alteration serves all of the following purposes: (a) it doubles the Hit Dice in all the categories from Fine to Large; (b) it brings the Monster Manual average Hit Dice into the specified range for every size category (albeit near the top end); (c) it removes the anomalous jump at the Huge category, creating a smooth doubling at every step; and (d) it synchronizes the standard with the Elemental Hit Dice progression (leaving only Dragons with a unique methodology).
As a concluding note, it might also be suggested that the EHD calculation on MM p. 11 should have used the low-ends of the ranges above instead of what it originally listed (i.e., the same as the errata change for Elemental only) -- in fact, this would only require changes at the three categories of Medium, Large, and Huge (recommended EHD base of 2, 4, and 8, respectively). However, it is certainly impossible to think that the skills and feats for all the published monsters in these size categories could be revised or errata'd at this time; just make sure that as DM, you don't mistake the EHD numbers as representing "average" D&D monsters.
You may wish to look at a spreadsheet analysis of these two different Hit Dice progression systems, below. Therein, you'll see that the majority of monsters in the Monster Manual (dragons excluded) have Hit Dice in excess of the range specified for their type by Mr. Williams article -- nearly 60% in the critical sizes up to Large. Meanwhile, practically none fall to the low-side of his ranges*. On the other hand, my revised "smooth" progression finds that the majority of monsters (59%) correctly fall within the specified ranges, with a more equitable 13% below and 29% above the "norm".
I'm actually forced to conclude that this smooth progression was in fact used by one or more of the original designers, but that this behind-the-curtain system somehow didn't make its way to the desk of Skip "the Sage" Williams.
See this Microsoft Excel analysis, here.
* For more details on the anomalies in Mr. Williams' numbers, see here.