Since this provides a temporary benefit, it may make more sense to compare this to heal than to other DR shields. If we assume that all affected attacks will have at least 3|4 damage to lose, it is easy to calculate how much damage the shield will block on average.
Probability of breaking | Unupped(DR3) | Upped(DR4) |
5% | 60 | 80 |
7% | 43 | 57 |
10% | 30 | 40 |
12% | 25 | 33 |
15% | 20 | 27 |
20% | 15 | 20 |
First, I note that in its current form, it will block the same amount of damage when upgraded for the same cost (+the cost of upgrading, of course). Also remember that the damage blocked decreases anytime it is attacked by creatures with less attack than DR, which is of course more likely with higher DR. So, the upped form is currently weaker.
The shield costs 5
: + 1 card + 1 turn + temporary use of shield slot.
Heal costs 3
: + 1 card
The shield has the added advantage of blocking poison from scorpions for a bit, and the utility of being usable as a Shield (the unupped other card, cost 1 :other: + 1 card) until you need better protection or have a replacement ready. It has the disadvantages of being weak vs low attack creatures, vulnerability to PC, randomness, and the risk of being killed before it can do its job.
My instinct would be to have the probability of breaking in the 7-10% range, with the same value for unupped and upped forms.