Here's a proposal that aims to reduce an advantage that a selfish miner may have without a change to Bitcoin's block validation rules. The proposal suggests that as long as >50% hash power is slightly less willing to switch to the newest chaintip with the most proof of work -- at least, that's the best I got for understanding it.

the Bitcoin mining protocol proposed by Nakamoto (2008) and implemented in practice is well known not to constitute an equilibrium: Eyal and Sirer (2018) construct a profitable deviation called “selfish mining” which relies on strategically delaying disclosure of newly mined blocks rather than publishing them immediately. We propose inertial mining, a novel mining protocol. When miners follow inertial mining, they produce the outcome intended by Nakamoto, i.e., a single longest chain. But unlike the Bitcoin mining protocol, inertial mining constitutes an equilibrium (assuming no miner controls more than half of the mining power). Indeed, neither selfish mining nor any other deviation is profitable. Furthermore, inertial mining only changes miners’ behavior in the event of off-path forks, and can be implemented in Bitcoin without any changes to its consensus mechanism or blockchain architecture.

I think the paper may overstate the power of selfish mining:

The seminal paper by Eyal and Sirer (2018), however, establishes that the standard protocol is not an equilibrium, and constructs a profitable deviation strategy they call “selfish mining”.

But, they propose a different type of mining strategy that as long as it was adopted by >50% of the hash power would reduce the benefit of a selfish miner.

The inertial mining protocol we propose works as follows. Consider miners who are working on the last block in the current longest chain. If a new block gets appended to this chain, the miners will switch to working on the new last block, as in the standard mining protocol. However, if a new chain is published that does not extend the current longest chain, the miners will only switch to it if it is longer by at least I>0 blocks. The number I is a parameter of the inertial mining protocol. To ensure that a specific symmetric strategy profile is an equilibrium, I needs to be chosen to be large enough, as a function of the miners’ distribution of mining power. The closer the mining power of the largest miner is to half, the larger I needs to be. When there is a miner with power half or more, no choice of I constitutes an equilibrium.

Inertial mining differs from the standard mining protocol in its prescription of which chain to append to if there is more than one public chain, but results in one single chain as equilibrium outcome and thus achieves the intended outcome of the standard mining. Importantly, it does this robustly as an equilibrium, with miners having no incentive to deviate. It is straightforward to see that under inertial mining, selfish mining is no longer a profitable deviation. The main technical contribution of this paper is to show that no other possible deviation is profitable. This is challenging because the set of possible deviations is large. Our proof addresses this by assigning each honest block that is displaced from the equilibrium chain to a particular strategically mined “killer” block, and then showing that no such block can be expected to displace enough honest blocks to outweigh the miner’s equilibrium share.

tl:drtl:dr

Under normal mining:

The selfish miner reveals B+B1.
Everyone sees B chain is longer.
Honest miners switch to mining on B.
Some honest work on A gets wasted.

Under inertial mining:

The selfish miner reveals a rival chaintip that is B+B1.
Honest miners do not necessarily switch to mining on B1.
They keep mining on A.
The selfish miner now needs a larger lead before honest miners abandon A.

The paper does not identify what the larger lead needs to be, but does attempt to demonstrate that there is such a value that would change the incentives of mining such that selfish mining is not profitable.