Why use throw in modifiers instead of conditional entrance?

Question

Almost every tutorial advises a code like this for a permission modifier.

modifier isAdmin() {
    if(msg.sender != owner) {
        throw;
    }
    _
}

But wouldn't it be simpler to use it without throw statement (AFAIU it would also save user's gas):

modifier isAdmin() {
    if(msg.sender == owner) _
}

As it is described by ConsenSys in Ethereum Contract Security Techniques and Tips but only for these modifiers:

modifier stopInEmergency { if (!stopped) _ }
modifier onlyInEmergency { if (stopped) _ }

Answer 1

If you send some ether in the transaction, it gets returned to the sender when there is a throw (see this). In your example, if you sent some funds (in msg.value), and the modifier condition is not met, your money is gone, but the code after the modifier is not run (and if the function was supposed to return something, it would return only the default values for parameters, so you might not even know something bad has happened).

You are correct about the gas, in case there is a throw, all gas is consumed.