Backgammon is part luck as far as the dice go. However what you do with the dice rolls is the skill part. The players with better records and ratings are the ones who know how to use the dice rolls to their best advantage. IS there luck involved? Of course. However there has to be some skill as well or how else would you explain why some have great records while some have bad records. Nonetheless, if your primary objective is to have fun then what does any of the ranking stuff matter.

Backgammon is a game of both skill and luck. It's far more skill than people realize. And just like the varying skill levels in any game, Backgammon has its own varying skill levels. Can a beginner beat an experienced player? Yes. Is this totally dependent on the roll of the dice? No. Poorly placed pieces regardless of the roll of the dice can almost guarantee you a loss. But more importantly, an experienced player will win the majority of the games against less skilled opponents (statistically speaking).

I don't know if the debate here is the luck of the dice and how that places an unfavorable advantage to the lower rated players as far as the ratings setup here goes, or if the debate is simply over the issue of luck vs. skill.

Here are some sites to check out.

Salar's site
Backgammon
Newsgroup Archive
Skill vrs Luck
Backgammon
balances luck with skill
A Measure of Luck
Various articles

My own opinion is that two players of equal strength playing at their best will be at the mercy of the dice roll at least at some level. However, since there have been world champions who consistently win even against highly skilled players, luck must be statistically ruled out.

A higher rated player has a lot more chance of losing to a lower rated player in 1pt games. In a 1pt game a gammon or backgammon means nothing to the end result of the game, also a short game is much more at the whim of the luck of the dice. This is why you will rarely see the top players playing 1pt games at dailygammon or games grid, or playing them with a separate ID so as not to affect their ratings.
In multipoint games the odds even out a lot more, a higher rated player has a lot more chance of beating the lower rated player simply because of game skill and strategy.
With FIBS the lower rated player is rewarded for beating a higher rated player, but the higher rated player does not lose such a disportionate amount of points.
Here on the other hand we will see a similar thing happen as what happened at Goldtoken.
The top 20 players consisted of quite a few players (most) who would refuse a casual invite, most were non club members thus were not involved in random match ups in club challenges. Some only played 1 or 2 other players. It also allowed a player with only 35% (or was it 33%) wins to achieve the number 2 position simply because the games they won were against the number 1 player, who just happened to be a buddy. When the system was switched to FIBS and all games were recalculated some interesting things happened, players who had good win/loss ratios but had previously maintained a low rating simply because they played anyone, had a large rise in rating, whereas other players who had a high rating all of a sudden took a big drop. FIBS rewards players for winning, you will not simply get the large point gains by winning the occasional game against a high rated player...nor if you lose will you see your rating drop by 30-40 points at a time, only to find your next win is rewarded with 6.
Backgammon is a very different game to chess, it is extrememly popular, and is usually one of, if not the most popular game on any game site. Thus I personally feel it deserves to have the most effective rating system applied, this will become even more important if multipoint games are ever introduced to this site.