I think it would work in more situations and wouldn’t would be as hard to program as some might think.
Micah, do you realize you are talking to the person who wrote WinTD and thus has a VERY good idea of how hard something is to program in it?
A bad idea that’s easy to program remains a bad idea.
and a good idea that that’s hard to program (but not as hard as some might think) remains a good idea.
In all candor, Mr. Smith, intransigence such as this is precisely why you get the response you do on so many things.
You propose something that involves a programming change to WinTD. The author of WinTD (who also happens to be an experienced tournament director) gives you direct feedback on your idea, explaining why it isn’t good. Yet, you wish to continue to argue with him.
If you won’t listen to Tom Doan himself about the program he developed, who WOULD you listen to?
Why bother listening to someone with more expertise than oneself? Think of the time it would take away from the constant creation of all these new, great, ideas! 
Instead, it makes more sense just to put them on the foes list.
Questions in this post are to be considered rhetorical, so you can understand that apparently simple ideas often aren’t.
So even you recognize that this would largely memorialize the first round randomization. Where it doesn’t, it would often lead to decisions which are actually counter to common sense—suppose that one of the players gets a downfloat to get crushed by the highest ranked player below. His CA goes up substantially relative to others, and he is rewarded for that with tougher pairings subsequently. Players who get downfloated generally aren’t happy about it to start, and to have a pairing system which (by design) keeps feeding them harder pairings than others similarly situated isn’t a good one. So in that situation it’s worse than simply sticking with the original randomization.
There are also problems with implementing it:
how do you handle byes and forfeits?
how do you handle games involving unrateds?
The best you can come up with is that there are “some instances” where it might be better. OK. Supposed player A goes 1600W, 2000L and B goes 1610L, 1090W. Certainly it would appear that A is a stronger player than B. However, you’re not now making A an 1800 for pairing purposes and B a 1350. A will be nnnn(1) and B will be nnnn(2). If they are on different color schedules, then the (1) and (2) don’t matter, color does. The rank ordering within the group only matters if there is more than one player in the group with the same color schedule. So the pairing program would apparently have to locate situations where the rank ordering would actually matter, prompt the TD with the players’s records, ask her to rank order the players, which might not be a cut-and-dried decision. Suppose the TD had seen the two games and saw that A’s opponent overlooked a cheapo mate and B flagged in a dead-won position. And the result of this—in this case, it would be which of A and B gets paired with the 1570 and which with the 1540. Guess what. No TD would want to do that.
So the proposal could best be described as “no gain, just a pain.”
That doesn’t mean he is always right and you can never discuss (not argue) your ideas with him further.
Yes, in this situation it wouldn’t be a good to differentiate based on performance rating.
The program would put the player with the higher rating performance first, so their would be no prompting the TD to rank the players.
When it comes to the program he wrote, actually, that does mean he’s pretty much always right.
Again, there is an unanswered question. If you won’t listen to Tom Doan about WinTD, who WOULD you listen to? Based on your history in the Forums, I suspect we already know the answer to that question.
Except when someone lists about a half-dozen flaws with your proposal, and you ignore most and pick trivial nits with others, you’re not discussing, you’re arguing.
Somehow, someway, the TD would have to be presented with the list of how the players would be ordered by CA, and make a judgment call about whether to accept that, alter it or ignore it. Again, no TD would ever want to do that over such a trivial difference in the pairings. It’s a bad idea. Period. Let it go.
Well, no TD who was any good would ever want to do that. But then we generally understand that there are many legal pairings that are possible, and sometimes the "best " pairing is a matter for disagreement as such determination is subjective. But then, again, we tend to understand the purpose for and application of the pairing rules.
Why? Just go with the ranking by performance rating under the scenarios where it is usually appropriate to do so. Their are other pairing situations that it would be best to have to the TD occasionally make a judgement call on as well but as this is obviously not going to happen, we go with what the pairing program gives us.
Only incompetent TDs do that. The (major) pairing programs do a very good job, however there are times when it makes sense to tweak the output. I did that just yesterday, in fact.
Alex Relyea
Having them do a very good job at pairing is predicated on the TD doing a very good job on set-up. Most of the time when I have to tweak the pairings I investigate further and find that I actually have to tweak the set-up and the pairings will come out the way I think they should have in the first place. That said, I do have to occasionally override what was done even after verifying the settings worked well. I did that for tournaments on 12/30/2012 (avoiding a round three trap in a 6-player 5-round Swiss - though later I realized that setting it up as a round-robin directed Swiss would have probably been better) and 2/2/2013 (shifting the bye off of an unrated player in a 4-round Swiss even though it was set to not drop the unrated player).
Point well-taken. One can never overestimate the boneheadedness of at least some TD’s. One advantage of writing the pairing software is that you can at least steer everyone clear of ideas that have no redeeming social value.
Given that, I suggest you write forum posting software.