Lucius Marcellus wrote:Gremrod, introducing the Research Pool system (which is really just allowing full specialisation), will just allow people to avoid 70%+ of the research, which are 'low demand items'. So this effectively makes the grind a lot shorter (who in their right mind would research Cargo scanner?).
That would be true, unless there was a exponential penalty appiled if the delta between item research branche (tiers) becomes too high.
Lets say that a branch of the tree has (30) modules.
T1 kernel is worth 1 point, T2 kernerls 10 points, T3 100, and T4 1000. (for example)
Base cost for learning a module is 100 points for T1, 1000 points for T2, 10000 points for T3, and 100000 points for T4.
Now that looks like it would be fairly quick to get from T1 to T4 of any item... BUT
There is an additional multiplier for every unlearned module in the previous tiers and a penalty for each known module in the current tier.
What this does is simulate the effect of 'lost' kernels.
So, in the example of 30 modules in say the light/assault tree. There is no T0 research, so your 'previous penalty' (PP) would be zero. You have no current learning, so your 'current penalty' (CP) is also zero.
That means you can learn a T1 module for 100 points. Very quick, very simple.
Learning T1 unlocks T2 of that module. But now penalties apply.
PP (prevous tier penalty) is total - known; 30-1 = 29
Learning T2 (same module) CP=0 at T2; is 1000*29 = 29k points. Not too bad.
PP = PP(T1) * PP(T2) or 29*29 = 841
Learning T3 (same module) CP=0 at T3 is 10k*841 = 841K points for T3
... in perspecitve that's 8430 T3 kernels or 840 T4 kernels worth of points. your going to really have to 'WANT' that now to spend that.
But you can now see that T4 is going to be 100K*29*29*29 about 2.4 M points.
however, that only gets expensive like that if you try to specalize to T4 on 1 item. The ability to do it is there, but you do so at great expense.
Look at T1 again. It takes 100 pts for the first item 200 for the second, etc until 2900 for the last item (100*CP). Which I believe is only 465K to learn all of T1 items.
Once all T1 was learned, your PP is 0, so the first T2 item would simply be 1000 pts.
----
The 'trick' here however is when you add in new items. Unlike the 'random' method, the player knows exactly which tree and branch to apply points to when the new item is introduced.
However, it would still be very expenisve for beginning researchers to learn a new item too high in the tier, while vetern producers would still have to pay a stiff CP cost.
The non-random method works just fine, it can allow beginner researchers to do low-tier specializations, while still requiring a significant amount of kernel hunting for high tiered items.
The values of the kernels may need to be tweaked a little, but you get the general idea of divergent math per level and compounded penalties per tier.