For better or worse, being unusually concerned about having one's precious ideas stolen is one of the classical hallmarks of a mathematical crank. Therefore if you get visibly protective, the loss of reputation you'll suffer simply because you sound like a crank is likely to outweigh, by far, the very minor risk that you lose the opportunity to earn some reputation because someone took one of your ideas and ran with it.
Part of the equation is that just about everyone who actually does new mathematics always seem to stress how much more productive it makes them to discuss their ideas with an audience. If you can find an audience who are capable enough to be able to make something with your ideas if they (hypothetically) did take them and run with them, chances are overwhelming that they will have too many ideas of their own to give all of them their due. But there's a good possibility that some of them would give you a few minutes of their time to ask the fortuitous question that pushes you in the direction of the eventual solution.
And even if your presentation inspires someone to do some real work on top of yours, it is overwhelmingly more likely that they'll suggest a collaboration than just appropriate your ideas as their own. Doing so is not only the right thing do do -- it is practically without cost to them, because being known as "Jones who published such and such" is not much more prestigious than being known "Jones, of Jones and Smith (who famously proved such-and-such)".
Also, if your ideas are that good, they'll want to stay friendly with you so you'll let them know about your next good idea and give them a chance to collaborate on that, too.