If this doesn't stack, it should be cheaper.
Glad you're taking my shield coating to heart 
Stacking will increase the chance of the effect occuring, a portion of the chance may also be related to the damage that would be absorbed by the shield.
Having experience playing with and against SoFr decks, I think 25% may be a bit too much... But 10% might indeed be to low.
I think 17% would work well. That way with a full stack of 6 it caps out at ~100% like it should.
Does this seem good enough or would it still be to low?
It should be noted that if it goes too high too fast, some shields (like hope) will create OP combos.
I'm intentionally trying to keep it slightly less powerful than SoFr so that it is subtle enough to remain as a type "Other" card.
I recently posted a topic with a similar critique of SoFree: Expanding Evasion Mechanics
Anything with 100% defense is going to be OP if it has any significant place in the game. The same potentially applies to your card, which is -- I'm guessing -- why you made it <17% to begin with. However your card has the "advantage" that removing the shield makes the whole stack useless, as well as the potential for it to be UP for 1-point or lower shields.
Possible solution:
Chance to protect allies is 10% + 24%/R, where R is the damage reduction of the shield.
Dimensional Shield, Bone Wall, Wings, Dusk, and Fog all count as R=infinity.
Chance per Insignia:
Infinite: 10% + 0% = 10% (6 stacks is 60%)
(3-point) = 10% + 8% = 18% (6 stacks is 100%)
(2-point) = 10% +12% = 22% (5 stacks is 100%)
(1-point) = 10% +24% = 34% (3 stacks is 100%)
However this still offers only ~.3 - .5 reduction per Insignia, so I would probably make the cost 2|1 
Would this wording work here:
"Gives your shield a chance to protect allies, based on stack size and effective protection. All damage treated as physical."
or is it too vague and likely to leave players confused?
I would use the formula:
-- Stacksize*10% + 40% / R
Where R is the effective damage reduction
For most shields its just damage reduced.
Shields that deal a special effect that is inapplicable get +1 damage reduction to account for the effect boost
e.g.
Frost shield, procrastination, skull shield, thorns, fire shield, all get +1 to damage reduction
Shields that cause a miss chance or full absorb have R calculated as damage being dealt times chance of a miss or percent absorbed / blocked.
-fog, dusk mantle, dissipation shield, bone wall, and dim shield.
Weight shield and wings would be special cases. For spell damage sources, they are treated as 1 point reductions. For physical damage sources (e.g. gravity pull) they are treated as full damage block or 1 point reduction, whichever is more.
Mirror Shield and Emerald shield just get +1 damage reduction instead of reflecting back all spell damage. This would be far too powerful otherwise, since most damage that is dealt to creatures tends to be spell damage.
I think that covers all shields. In every case, a shield would be able to provide at least 1 point of damage reduction to allies.
