Ok, here's my idea.
While a 100M league is not a "low cap" league, the need to extract value out of every dollar is higher than it was at 120M. I took Tris Speaker last time because, even though his second most expensive season is not a good value, 120M gives me enough flexibility to make up for it elsewhere by taking a hyper-value player (I believe I chose to do this in the bullpen, primarily). There are very few players whose two most expensive seasons are also both a good value from a cost-per-win-above-replacement perspective. Some notable ones who were selected are Jimmy Sheckard, who I considered taking both this time and last time, and Benny Kauff and Andy Rincon, who were taken last time. Cy Young and Carl Hubbell are also likely good picks at this cap, along with a few others.
Additionally, at a lower cap, the differentiation between Player A vs Player B is smaller, so the benefit of exclusivity is lower. So far, 3-5 players I'd consider using to fill out my roster with have been taken but, since it's at 100M, there are replacements who are equally good (if not better) values. 3-5 is also pretty low considering 46 players are chosen in this draft that I can't use, but that's more likely to happen at a cap like 100M where the useable player pool people are considering when making their picks is much larger than at higher caps. This ends up creating no incentive for me to worry about exclusivity, and tilts my decision making only towards extracting value from my roster.
Especially at 100M, spending 10% of your entire team's salary on less than 10% of your innings (Kimbrel, Jansen, etc.) or spending 15-20% of your entire team's salary on 25% of your lineup (Pujols, Kiner, etc.) scares me because I would have to sacrifice significant value elsewhere to fit under the cap. My hypothesis is that the sacrificed value will be higher than the value I would gain by having those players over using the leftovers from the draft.
To be fair, I normally wouldn't roster 2 mopup pitchers, so I do lose some Effective Salary by spending an additional ~220k in dead money on the more expensive Cody Martin season. Even so, I probably would normally spend that much additional cash on my pinch hitters, so I'll be able to compensate by getting a bit cheaper with my bench.
With all of this said, because (I feel) this strategy is likely to be a winning strategy, I would recommend either a higher cap or removing the "two most expensive seasons" piece, and implementing a minimum salary spent on drafted players for future seasons (likely for both hitters and pitchers, individually) regardless. But let's see how it plays out in the league.