disputes: punishment on repeated dispute initiations (stale)

[ordian]

As of paritytech/polkadot#5535, we don't disable validators for "against valid" disputes. But based on the number of disputes initiated by a validator within a session, we might want to disable it if it reaches a certain threshold. Since submitting more such disputes wastes the network bandwidth "for free".

We can also consider doing exponential slashing or slashing up to a certain number or what's suggested by @burdges in #790.

But disabling a validator would have a greater short-term impact since slashes are only applied after a bonding period.

[tdimitrov]

Considering #784, is this still relevant?
We ironed out the disabling part but what about the exponential slashing>

[Overkillus]

Will be solved by the new disabling strategy. 0% slashes dispensed for validators voting invalid on valid will cause disablement. Disablement has opportunity cost which should be enough of deterrent.