Delta and 3-phase Jerry-rig produce the same (unloaded) output voltage. The difference is in the current and resistive losses.
At any given moment one of the three windings has the highest induced voltage. If the waveforms are true sine (or a number of others that add up right) the induced voltages of the other two windings total to this same voltage.
Both Delta and Jerry3 source the same current, and have the same resistive losses, from the high-voltage winding. The difference is in what happens in the other two.
Delta sources another 50% of the current from the other two windings. (Voltage the same, series resistance doubled.) With half the current through twice the resistance the losses are in proportion.
The current sourced by Jerry is more complicated and depends on details of the waveform, the loading, and the instantaneous phase. During part of the cycle (when most of the voltage in the off-windings is across one of them) Jerry-rig produces a HIGHER current from the off-phase coils, because it only sees one coil's resistive drop, not two. During other parts (when things are more evenly divided) the sum may be lower, or even zero. (And this will depend on loading, too. High enough loading that more than half your voltage is lost in resistive drop in the coils for a delta means the equivalent Jerry rig will still be sourcing current even at the 50/50 voltage split.) Resistive losses for a given current will be half as much as with the off-side of a delta, because they only go through one coil's resistance rather than two. But because the current is more concentrated into a portion of the waveform and losses go with the square of the current the average efficiency over a cycle will not see all that improvement.
Further, if the waveforms aren't ideal a delta will have circulating currents. Even with small deviations from ideal waveforms these currents, and the losses from them, can be large. Jerry rig doesn't have those losses.
So while I would expect more output from delta than jerry for a genny with a near-sine waveform, I wouldn't be able to confirm it without some tough math and some assumptions on operating conditions - and it would be reasonable for Jerry to be close or maybe even do better. But for the waveform-of-opportunity gennies we build here having Jerry-rig come in significantly ahead is very reasonable. B-)