1: Leave it as it is.
2: Make a scoring cap.
3: Compromise: make an even higher score less effective (like using a square root)
Whatever scoring system we use, most good players will always alter their game play to get a high score, so Starlifter, your argument that a cap will limit playing styles is true, but not as relevant as you think because any scoring system does. You admitted in another thread that your game play has changed toward early growth due to the GOTM scoring formula. If you're saying you don't like the WAY it limits playing styles, well OK.
Some analysis on suggestion 3:
Let the GOTM score at time t be g(t).
Let the ordinary Civ score at time t be x(t).
Let T be the total number of game turns.
An increasing GOTM score means g(t) > g(t-1) which is equivalent to the ratio
x(t) / x(t-1) > 50^(1/T).
The right hand side is largest when T = 420 and = 1.00936, so the criterion for continuing farming instead of launching is that growth in your normal score is at least about 1% per turn.
Changing the formula to use sqrt(x(t)) means that the GOTM score will increase if
x(t) / x(t-1) > 50^(2/T).
This makes the criterion change so that normal score growth should be at least 2% per turn. This is harder to do later in the game as more cities max out, so will make people launch earlier if they can. The change in strategy will be to found a lot of cities all at once and grow them at the same time. Game play will be more "phase" based. If the top scoring players (who are the only ones greatly affected by these changes) don't mind this strategy change, then vote for it.
If you want the best GOTM score under system 2 or 3, game play won't be quite as different as many people seem to think. The advantage to system 3 is that we keep the high score medal, so I think it's a bit better, also because it allows more flexibility in game play than system 2.
In an earlier post in the thread "GOTM scoring formula", I said that we will always have problems and arguments unless a scoring formula is DERIVED from a set of principles on which we try to agree first. Suggestion 3 still won't help the early conquerors or SS landers who don't use WLTPD much.
Matrix, I know you said to keep it short, but I think the analysis here is valuable.