C = Copper (gold/bronze)
E = Ebony (black)
G = Green
P = Purple
R = Red
Y = Yellow
Algebraic notation (similar to chess): Columns left to right are lettered abcdefghi. Rows bottom to top are numbered 123456789. For example, the bottom left square of the board is a1.
I went ahead and analyzed this Pro game of 504. I recommend you read through my comments as you look through the game. There were no big mistakes from 0 to 70 (of 504).
At 70, after cashing the R column, Ch1-e8 is bad, because it will take you two more moves to poise the 8-row. Instead, you should create a C three-line diagonal (e.g., Cg8-c7). Less attractive would be to create a G three-column (Gc3-a5) because it is blocked by Ya3, but that would still be much better than Ch1-e8 as played.
[An efficient move changes your ball mass by -2.5 (negative numbers being good). Because you’re fighting a +3 drop each time you make a move (cashes don’t count), your ball mass increases by 3 – 2.5 = 0.5 per move (which I don’t count in most assessments, it’s just a fixed constant that you’re fighting). To break even over a sequence of moves, you need a turn-gaining drop 20% of the time, or actually more often when you account for the fact that you will occasionally be forced to make an inefficient move.]
Ch1-e8 (as played) only helps your ball mass by -1 (instead of the standard -2.5). In short, it is an error of 1.5. Looked at another way, just prior to cashing that line, you will have invested three moves (value of 7.5) to poise the removal of 6 balls (thus a mass increase of +1.5). By contrast, Cg8-c7 will require only two moves (this one plus one more) to poise the removal of 5 balls, thereby meeting the standard of -2.5 per move, and maintaining an increase of zero.
At 92, after cashing the E diagonal, Pc2-c4 is a blunder of +2.5. Better is to play Gc3-e8 or Gc3-a7. Even playing a Y to a4 (if legal) would be much better than Pc2-c4.
At 152, Ya3-g5 is a blunder of 2.5. To hold par, you need to move a Y to h7, creating a 4-line column. (Oversight?) Even extending a C or G two-line (or some independent Y two-line) would be better than Ya3-g5.
At 192 (after you played Rg4-d3 and Rf3 dropped), moving Ge3 is a +2.5 blunder (there is positional compensation in unblocking the central R row, but I’d estimate you get back -0.7 at most). You should keep busy elsewhere until the drop gives you a G two-line somewhere. For example, Cb7-a3 unblocks a B six-line.
At 244, it is a +1 mistake to cash with Yc1 because it is part of a potential six-line on the 1-row. You should cash with one of the Y balls in the upper left (or delaying the cash is a fine option).
At 254, after you poised the P row, you got an unfortunate Cd5 ball, but you should NOT move it immediately. Rather, best is to poise the Y column. For example, Yg2-i4/i5. Next Yg8 or Yd3 can cash, followed by (a) Gc4-g2 + Cd5-a3, or (b) Bc2-i6 + Eb2-e6/g8 + Cd5-a3.
Granted, both plans might get blocked, but some resource might arise along the way that could be a back-up (e.g., a C drop pairs somewhere) or even an improvement (G drop on the 4-row could open the Cd5-a3 path faster while gaining an additional turn).
As 294, after you cashed the B column, Yb6-e4 is a blunder of +2.5 (actually 2.0 if you are able to cash the e-column as an 8-line). Better is Yb7-f2, so that you can cash two five-lines instead of one. Examine this carefully.
[Scroll both diagram forward two frames, by clicking twice on the fourth button ]
The Yf1-e1* actually misses an opportunity to save 0.5 (as I hinted at parenthetically in the previous paragraph). I like Re6-a3, intending Ed7-e6, E poise, E* + Ce7-c7, Y-e7 and Y-e6*. If it doesn’t work out, I still have Yf1-e1* in my back pocket. No big deal, though; the real mistake was Yb6-e4.
At 394, after the Yh6-i5 poise (which was fine), the Y* was a hard-to-see mistake, costing +1. You can get -2.5 for the 2-row and eventually -3.5 for the i-column (including the Yi8 ball) for a total of -6 (compared to the instant -5 as played).
Hence, what you should do is use Cd2 and Ch8 to cash the C diagonal, and move Ri7-a5 (though no hurry). The next two Y drops can be used to cash the 2-row, and Yh1 can be used for the i-column later. You can even get by with one Y drop if you move Gf1 somewhere (e.g., f8), thereby liberating Yg1. There is no hurry, though, on the Y plan. You might just rather play elsewhere for the next dozen or two moves.
At 494, Ef2-d5 is a big mistake of +1.5 (2.5 less than ball unit you get back for the potential vertical six-line). Best looks to be Yh4-b1, or Yb7-i5 may be better yet if an E is on deck to drop.
[Scroll both diagram forward six frames, by clicking six times on the fourth button ]
At 494, after the E poise, the E* costs +1. Instead, I would wait for a P pairing. If an E (but no P, G or R) is about to drop, then Eg5-e3 looks good, followed by cashing the E row. The E six-line potential remains, but you buy more time and space for a P drop to pair somewhere. If no E, P, G or R is about to drop (or you don’t know), then Ra7-e8 looks good, unblocking the Ga4+a8 doublet.
You also made all kinds of unnecessary minor errors. In particular, you should strongly resist breaking up a healthy doublet. Instead, find a different source ball, move a different color, or DELAY THE CASH. (Admittedly, delaying the cash makes the resulting position harder to visualize, which in the short run may lead to more mistakes, but eventually you’ll discover that delaying the cash has many hidden benefits.)