No - re-read what I wrote. If gray is there, that horse WILL be gray.
The 50/50 is based on what that horse can pass to its offspring. If the horse is heterozygous gray, Gg, then each breeding there is a 50/50 chance the foal will be gray (again assuming a non-gray other parent). Every gene is 2 copies. Every breeding gives a 50/50 chance one or the other of those 2 copies will be passed on.
And not all grey horses turn “white”.
If they live long enough they certainly do
If there is a grey parent, not all foals will be grey.
Never said that - please go back and re-read 
However, if the gray parent is homozygous gray GG, then yes, ALL foals will be gray, regardless of the gray/non-gray status of the other parent. That’s because the only copy of that gene to pass on is G, the “on” for gray, and when G is present, the foal WILL be gray.
The color on my arab gelding was very strong on the sire side.
What does that mean? His sire was chestnut - absolutely no chance of gray coming from that side. His dam was gray. HER sire was not gray, the dam was. That makes her Gg, heterozygous gray. That means every breeding gives her a 50/50 shot of passing on the G (which would produce a gray foal) or the g (which means a non-gray foal). Your guy got her g, therefore he is not gray.
His sire’s chestnut “strong color” did not overcome the gray. There WAS no gray given to the foal. The stallion had absolutely nothing to do with the odds of the gray dam passing her G or her g.
If I had another grey mare, I would breed to a horse who had zero or very little grey in the pedigree, or when bred to grey, didn’t produce it.
It doesn’t MATTER what’s in the pedigree if the horse isn’t gray.
If the horse IS gray, then you keep looking back. You can have a line FULL of grays and them ALL be heterozygous gray Gg and bam out pops a non-gray foal.
Still, regardless, it is up to genetics to determine the color. I dislike grey, I have had many many, so I will not be buying or breeding to any grey ever again.
Yes, it IS up to the genetics, and I and simkie and Daventry and quicksilver have explained several times how it works.
Greys in the egyptian lines are very strong. Not so in say the polish, bays are. In the spanish there are greys and bays usually. Russian, chestnut is very dominant. Crabbet, chestnut.
ALL that means is there are more homozgyous grays in those lines/that breed. Look at Welsh ponies - HUGELY gray population, but there are some non-grays. MUCH more likely to find GG grays in Arabs and Welshes. But that doesn’t matter. All that matters is the horses that are bred together. If 2 arabs are not gray, then it doesn’t matter if their 10 gen pedigree is gray, they will not produce a gray foal when bred together.
But like I said above, if I had one to breed I would choose a color and stallion who had a VERY strong color dominance, that is if you wanted the best chance to not have grey.
If you don’t want gray, don’t breed a gray horse, don’t breed TO a gray horse. Nothing else matters, not the horse being bay or black or chestnut, not even if the horse comes from 10 generations of black or 10 generations of chestnut. If one parent is gray, there IS at least a 50% chance of gray, regardless.
I do think it does have to do with the breed. Grey is not a common color in TBs. Arabs, it is. More chance for an arab to be grey, than a TB.
it DOESN’T have to do with breed.
Gray is gray. Black is black. Bay is bay. Horse color genetics are the SAME. Gray in a TB is not any different from gray in a Welsh or Arab or Morgan or Shire or any other breed or mix. It’s just not. It doesn’t matter what you think - the genetics are what the genetics are.