Lol, how much, would you say, as compared to running a major factory mine? 1/100000th? Less?
If you invest a million dollars into mining BTC, might you not throw a buck or two @ nodes?
Miners already run nodes. Why would miners be interested in mounting a Sybil attack that might break bitcoin into multiple ledgers? Seems that would be a poor business decision.
Heck, Bitcoin wiki tells me somebody was *incentivizing nodes* 'til recently:
Some are incentivizing it
Bitnodes is incentivizing full node operators "until the end of 2015 or until 10,000 nodes are running."[2] For rules and how to join the incentives program, visit Bitnodes Incentive Program.
Anyhow, does it matter how long the "bad" nodes stay up, if they're meaningful to Bitcoin security in any way (beyond being a good wallet)?
Sure, it matters how long attacking nodes stay up, because attackers are spending resources to do it. Like everything else in bitcoin, the system depends on incentives. The nature of an attack is to expend resources in the short term to realize bigger gains in the long term. That's why 5000-8000 Core nodes operating over the past two years are unlikely to be Sybils, and 800 Classic nodes spun up over the past week are more likely to be Sybils. The latter is bolstered by the fact that the majority of Classic nodes are "new" nodes, not existing nodes that switched from Core to Classic.
The "bad" nodes are harmful to security because they are run incompatible software (assuming they are not pseudonodes). So if miners do not agree on which chain fork is valid (in a contentious hard fork), these nodes will ensure that multiple blockchains survive.