Very curious myself why partridge would be suggested.
As far as Silkies goes, Partridge is about the closest pattern to "Wild Type" one could get(Think about RJF) other patterns have unnecessary genes that will get in the way, black silkies may have additional melanizers that may get in the way of the white crest. I would also suggest using a Partridge Crested Silkie
As for hatching 50 chicks. IMO that will be no where near enough chicks. I would hatch a couple hundred if it was my project.
You have a lot more genes that need to fall in place besides the silkie feathers and WCB color if you want anything that actually resembles a silkie.
Guessing it would also take a few more generations to get things locked into place.
I agree with that, 50 is a very small amount to lick, WC, 5 toes, silkie feathers when crossing F2xF2.
But now I advise against doing a F2xF2 cross and instead, breeding back to crested silkie parent line, I have confirmed that the white crest phenotype is a dominant trait and one would be able to pick up the BC1s chicks with Extended Black and White Crest from the partridge based BC1s.
So this is the breeding plan.
WCB Polish rooster over Crested Partridge(Or Black if you can't source partridge), You don't need many F1s 10-20 hatching eggs, just to make sure you get to select a very good F1 cockerel that showed the best white crest color at hatch(it will have a white chick crest down), take this male and cross it back to Silkie hens, at this stage I would recommend hatching as many eggs as you get as you would be looking for White Crest, Black Skin, Silkie feathered BC1s of both Genders, the best WC Silkie feathered hens crossed with the best WC silkie feathered cockerels..
Interesting Article about the study of Barring, Sex linked dominant dermal Inhibitor, the interesting part of it is that they used WCB Polish rooster over BR hens and could not tell the sex of the cross on the F1s and F2s due to the fact that the WCB genes were masking the effect(no comments on if it acted as completely dominant, but it seems that way)
https://pdfs.semanticscholar.org/3a3e/7a212b599853d965771a040b559e32dd324d.pdf