It's more like 10% increase on the outers. 100% increase with no spreaders. If the spreaders were drawn upswept, to subtend the 163 degree included angle between the vertical and the upper part, the shroud tension would be the same above and below the spreader, but I didn't want to work that hard.

But it does show the trend as the spreaders get shorter, which is all I was trying to illustrate.
I've stayed away from making any judgements about the suitability of shortening the spreaders since I don't know what the actual worst-case loads are, what the accepted loading profiles are for marine stays (as a function of breaking strength), and so on. There's too much industry specific stuff in there that I'm not aware of. But I was curious as to what happens as you shorten them, which is readily apparent from that graphic.
What IS rather interesting, though, is that regardless of the spreader length, or even the presence of the spreaders at all, the fixing point reactions are the same, i.e. the chain plates and the
mast base. The resultant force at the chainplates is the same magnitude (6.49F) and direction (81 deg) for ALL of the cases in the previous graphic, even with no spreader and just a single stay on each side, which is what you get effectively when you shorten the spreader to where it takes no load. 3.20F + 3.36F * sin(72.67) = 6.41F vertical and 3.36F * cos(72.67) = F horizontal. The square root of 6.41F^2 + F^2 is 6.49F, and the angle is arctan 6.49F/F is 81 degrees (rounding errors are in there, of course).
With no spreader, the inner stay also takes no load, at least in this simplified static point-load (at the top) scenario. The way it's all distributed gets lopsided of course, with more load on the outer stay and less on the inner (and the fittings and pins too, of course), plus larger compression load in a longer unstayed
mast length (with no inner stays), but the net result is the same at the deck.
Just a little fun with line drawings and hand calculator.
