In that scenario, I see no reason for even using J-16 instead of J-20.
The reason is the primary problem for PLAAF - training pipeline for pilots and ground crews. Militaries are groups of humans acting collectively and using tools - in that order - and wars are fought by units and not individuals. Once you understand that you see where the rationale comes from.
J-20 is a VLO airframe which really means it's a VLO in the horizontal plane, with emphasis on frontal aspect.
As with every VLO airframe you have to fly it in a very specific manner to minimise RCS and it requires relearning of all the flying habits and flying habits are trained to be instinctive so that pilots can perform in high stress and high g conditions. Flying must be like muscle memory or you can't do it in combat. Learning to fly a VLO fighter after flying a non-VLO fighter is like learning to talk again after a stroke. This is why it takes so long to introduce F-35 into active service. This is why for example when Poland ordered F-35s having already used F-16s for close to 20 years the US instructors told us to not even bother with sending young and promising F-16 pilots.
J-20 is also an unique design introduced into service only in the last decade which means the emphasis was on testing and development and not on mass introduction of ground crews.
J-16 has huge RCS so it doesn't matter how you fly and it is largely compatible with all the other Flankers in PLAAF service. With some 300 Su-27/30/35/J-11A/B in PLAAF service you gain a large pool of well understood aircraft with tactics - which involves cooperation with other Flanker pilots as well as other aircraft - that can be copied from one to the other. A J-16 is essentially a Su-30 with a more powerful radar. It is also simpler to retrain a technician with experience in servicing J-10s to service J-16s.
But training the units using J-20 and J-16 is one part of the story. Because J-20 is VLO and PLAAF is preparing for fight against enemy VLO fighters it will focus on developing both counter-VLO tactics for the entire air force, air defenses and the navy as well
as for developing collaborative tactics where J-20 works with other aircraft as part of a mixed composition.
There simply isn't enough J-20s available and won't be for the foreseeable future to completely disregard J-16 and make a leap from 4gen Flanker to 5gen J-20. Even USAF which had a 10+ year head-start with F-22s, and close to 20+ year head start with VLO in general is back-filling the gaps with upgrades to F-15s and F-16s.
People completely miscalculated how much of an organisational shift the transition from 4gen to 5gen would be and when the delays with F-35 began it effectively halted the intended introduction strategy. It's better to wait and train once on the intended configuration than constantly re-train while trying to force "concurrency".
Plots and technical ground crew are not the most intelligent people in society and they're being given some of the most sophisticated equipment currently in service. This is why AI-enabled unmanned systems are going to be such a quantum leap. It will turn the current dynamic which is constrained by human performance completely upside down.
But it doesn't mean that you can completely cut out the human element. EW is highly local. Intelligence even more so. If you have J-20s and twice as many UCAVs the disruption to decision nodes is easier than when you have twice as many decision nodes, even if some of them are not VLO. Relatively few missions in air warfare require VLO but all missions require intelligence.
In systems approach if you can isolate a combat or sensor node from its decision node you effectively neutralise it. Isolation i.e. disruption in communication e.g. with the use of EW is easier at longer range between nodes. If you have only the J-20s you can put your active assets only in so many places at so many times. If you add J-16s that number likely more than doubles, probably triples.
In the future neutralisation of decision nodes will be the primary aim of all air warfare because that's the primary aim of all warfare. So far air warfare was the exception because most of the time every aircraft was its own decision node to an extent. But if you look at ground warfare the dynamic is clear. Why waste resource on brute forcing the destruction of enemy combat potential when you can mission-kill it. That's how Battle of Atlantic was won - not by sinking German U-boots but by making them irrelevant as they had to stay in hiding.
J-16s are twin-seaters so they can manage greater numbers of UCAVs in defensive roles out of range of enemy EW disruption. "Intelligence victory" may be augmentation of a single pilot but also augmentation of a systems operator managing UAVs and UCAVs.
Mixing J-20s and other aircraft is hard. Mixing J-16 and other Flankers is easy. Which means that for many basic missions you can use J-16 as a lead fighter with other Flankers following. Sukhoi intended Su-30 in Russia's air defense force service with their "mini AWACS" system. Even on 1-1 basis with 200 datalinked J-11 and 200 J-16 you have in theory 200 pairs of aircraft more or less on J-16 level. If it's 1 J-16 with 3 J-11 then you can spare J-16 for other units as well.
PLAAF needs a reliable multirole workhorse with good ground strike capabilities and J-16 is just that. It's what Russia wishes it had because they're stuck with Su-35, Su-30SM and Su-34 instead of a single platform. Even USAF is choosing F-15E family as interim replacement of F-15Cs.
War is not a beauty contest.
Hopefully this clears your doubts.