I'm reminded of the fiasco that was the release of Privateer Press's army calculator for Warmachine. They managed to do real harm to their brand with that, and they're far larger and better established than Arcane Wonders.
Just pushing a poor quality something out through a contractor is one option, but getting it right is an investment. If 6 hobiests crank out six independent solutions, and one of them is good, all everyone remembers is that a hobiest solved the problem in his or her free time. If a company with established IP tries to cheap out and hire a bedroom programmer to put out a quick product, all they'll get is a 1/6 chance of success.
Magic the Gathering Online (for example) is a big undertaking by a serious group of professionals, and while Mage Wars would likely be simpler, it would still be significant labor to get right. I don't know if Alex has the rules nailed down so tightly that a program could handle all card interactions procedurally, but I do know that's a moving target anyway, that gets more difficult with each released set. Wizards of the Coast employs a full time rules manager and the full MtG rulebook is 192 pages long. That's the kind of specificity required to automate a game.
Rules are hard, is what I'm saying, and Arcane Wonders isn't a software company.
Software's hard too. Even simple seeming software is hard. Professionally project managed software is so routinely late and overbudget that the accepted method of setting the schedule for a project would be to plan out the time required for each part of the task and then to double it... except that when managers do that it still takes twice what they planned. In-house software development is no place for amateurs. I can think of few things more potentially ruinous to a small game business than independent software development.
I don't mean that the task would be impossible: there are some games coming from small studios of comparable complexity: Tyrant comes to mind. But possible and trivial are very, very, different things. It would be a major business undertaking, and I'm astonished a professional programmer would think otherwise. Even something as bare bones as OCTGN is passing the hat trying to get their network code rewritten.