While modern liberalism rejects realpolitik in theory, in practice the structures it has set up have often been ignored by purportedly liberal actors, so your argument from experience is very tenuous.
As far as prisoner's dilemmas, its a lot more complicated than that. A purely pragmatic approach can lead to win-win outcomes that would be impossible in a purely idealist framework - realist IR scholars would, for example, argue (and could back it up far better than the alternative) that if it wasn't for Bush and Cheney's idealism, the 2003 Iraq war would never have occurred.
The realist counterpoint to the prisoner's dilemma is that it's not just due to conflict prone relationships, but because it's impossible to know what the intentions of other actors are, which is not something that international liberalism can solve in practice. In theory, liberalists would argue that two liberal states can communicate their true intentions well because their intentions reflect the intentions of the population, but this has conclusively been proven to be untrue in the case of foreign policy (not very surprising due to the way intelligence has to be set up), and therefore the liberalist rebuttal to uncertainty of intentions seems much weaker now than it did originally.
As a result, realists argue that following realpolitik can actually reduce the likelihood of war, as agents will be careful to act in such a way as to avoid seeming as to pose a threat in the absence of knowledge of their intentions, while a liberal actor, acting according to their ideals, would act in a way such that an actor who is unsure of their true intentions would have to perceive as a threat.
This means that you just can't assert easily that realism is more likely to cause lose-lose outcomes. It's a complicated argument where this kind of ironclad certainty just isn't justifiable.
As far as prisoner's dilemmas, its a lot more complicated than that. A purely pragmatic approach can lead to win-win outcomes that would be impossible in a purely idealist framework - realist IR scholars would, for example, argue (and could back it up far better than the alternative) that if it wasn't for Bush and Cheney's idealism, the 2003 Iraq war would never have occurred.
The realist counterpoint to the prisoner's dilemma is that it's not just due to conflict prone relationships, but because it's impossible to know what the intentions of other actors are, which is not something that international liberalism can solve in practice. In theory, liberalists would argue that two liberal states can communicate their true intentions well because their intentions reflect the intentions of the population, but this has conclusively been proven to be untrue in the case of foreign policy (not very surprising due to the way intelligence has to be set up), and therefore the liberalist rebuttal to uncertainty of intentions seems much weaker now than it did originally.
As a result, realists argue that following realpolitik can actually reduce the likelihood of war, as agents will be careful to act in such a way as to avoid seeming as to pose a threat in the absence of knowledge of their intentions, while a liberal actor, acting according to their ideals, would act in a way such that an actor who is unsure of their true intentions would have to perceive as a threat.
This means that you just can't assert easily that realism is more likely to cause lose-lose outcomes. It's a complicated argument where this kind of ironclad certainty just isn't justifiable.