I agree, cheese should be defined as that line you cross when your mindset consists of "based on my own assessment of my and my opponent's skill, a situation i have control over is less likely to yield the desired results than a situation more reliant on dead luck", rather than the mindset of "given my opponent's specific skillset, the map pool, and/or the state of the mindgame we're playing, and/or, in favour of maintaining my reputation as unpredictable among players other than my opponent, an agressive strategy is the best strategy, even if it's extremely agressive." The thing that imo some people are missing is that starcraft is in part a gambling game, and you don't win in gambling games by taking exclusively soft risks. This is because you have to make a strong showing that you're willing to take that kind of risk, otherwise people will feel they can rely on you to act a certain way and all risks for them are reduced; the game essentially falls more into their control. This is precisely the kind of thing stork was trying to take advantage of when he delayed his second cannon so badly, despite not getting a clear scout.

The aspect that starcraft has that many pure gambling games do not, is of course that you always have the option of "playing solid" when you're so confident in the aspects which you can control that you can justify a zero risk scenario. However that concept works in the same way as the definition of "cheese" i've given. This would be the opposite of cheese, as in "with my assessment of my skill and my opponent's, i have a surer chance of winning at the clear disadvantage of playing safe all the time than i do by allowing any factors of build order luck the possibility of ending the game directly. That is, if i make it to mid game, i can turn around any situation." Jaedong might have assessed stork as being his equal, and this way of looking at things to be wrong. Or maybe he just enjoys putting it all on the line. In either case, so far his style has worked to his advantage.