Got busy and this project was nudged down the list by time and sleep deprivation, but it's still ongoing. With a Deadlands Classic: Hell On Earth game now joined, it seems a good place to examine things from for comparison. The company behind Savage Worlds I believe had little to do with Deadlands, but based their system on a much simplified version of it and purchased the rights to it, now sold as Deadlands Classic.
WHAT'S THE SAME?
Both Deadlands and Savage Worlds use D&D dice, from d4 through d12. Higher is better, and a target number system. Dice rolls explode in both systems if you score a max roll. Both use Edges and Hindrances to describe their system of advantages and disadvantages.
Deadlands has stats drawn from a deck of cards for random generation, distributed into stats. Skills are based on the 10 parent stats, each point of skill nets you another attribute-sized die to roll during tests, which you take the single best result of. Savage Worlds, skills are capped by attributes, but gain no further benefit beyond cheaper advancement. Instead, Savage Worlds grants player characters a Wild Die, which is just a d6 unless they have a few high level edges that boost its power in some or all situations. In SW, Skills level from d4 to d12 just as attributes do.
WHAT'S THE PROBLEM?
The problems with Savage Worlds are at least twofold, possibly more.
The first problem is that high stats give next to no bonus to skills based upon them. A person who minmaxes with a d12 in a skill-controlling stat (dexterity or intellect) will spend exactly the same amount of points to get a wide array of skills to average level as a purely average competing character with d6 stats across the board. For a character to be remotely competent, let alone skilled, they must stay fairly laser focused on a few skills to be good at (at least with default 15 skill points). The wild die allows a d4-skilled character to still have a d6 available, but it's still sub-optimal and leads to a number of character concepts that simply can't properly be made because they require too broad of skillsets (My first experience was trying to port in a weird west bounty hunter, who had ranged skill, melee skill, a vehicle skill, with tracking and perception. In GURPS or Deadlands Classic, such a build was very doable, low investment skills had decent power even without the law of averages to help out if the controlling stat was good.
The second problem is somewhat subjective, the lack of a bell curve. Single die rolls have a flat probability, SW Extras get flat probability and named characters get the luck die for a weird bell curve. Having a bell curve is nice because it gives you some semblance of idea how you can expect to perform on a given roll; ie, average.
WHAT CAN WE DO ABOUT IT?
There's a few possible solutions. A simple one could just be to give more skill points starting. Giving everyone at least the same bell curve could be done by giving Extras an Extra Die, perhaps a d4 because fate doesn't give as much of a shit about them as the Wildcards, but still maybe a little. But that still doesn't solve the problem of attributes vs. skills disconnect.
What you could do is replace the wild die with an attribute die. Suddenly a high dex character can bring their natural aptitude to bear with only a little specific skill, while a high dex/high skill character is a true terror. Or, add the attribute die in addition to the wild die, or offer choice of best two out of three. Offering this same upgrade to the NPCs would be powerful, as they're not necessarily built the same way as players, and tend to just automatically have the required skills/stats they should for whatever place they'll be encountered in.
I think that's enough rambling on this for a single post, I'll finish it up with actual probability calculations and such in another post.