Magic The Gathering: Arena Dev Tracker

So Arena actually reads the literal card text, rather than you translating the card text into a machine-friendly set of rules?

If by "you translating", you mean directly coding cards, this is correct.

That's super interesting, any chance you could shed some light (or have written elsewhere) on what made you decide to do it that way?

With 15 years of development on card games, it's not too hard to imagine in incremental improvements-

• Manually implement cards as code. (MTGO early days)

You get tired of doing boring work, so you move up to

• Template and generalize some of the manual coding up to a point with Regex + code generation. (MTGO modern days)

And that handles things like "Deal 1 damage to any target" or "Deal 2 damage to any target" just fine, but it doesn't handle stuff like "Deal X damage to any target, where X is the amount of damage dealt...

Sent :)

It's funny you mention this, because Underworld Breach hit our team like a bus, impacting everything from designing how we order action costs, to new tech in card representation on the client (turns out that we had never granted an ability with a variable cost before). We had never attached an ability from one card to another card based on qualities of that card!

Nevertheless, it's always about how complex the english is, not how complex the ideal implementation or backing rules are. "Escape {r}" is WAY simpler than Scourge's text. For escape, we literally just wrote code that explains what it means. For Scourge, we have to translate the english.

To add to what Aaron's saying in this thread, our time budget for JMP was pretty limited. We focused most of the effort on the novel cards in the set and those that are pioneer-legal. #wotc_staff

But what about Demonic Embrace from M21? It is the newest of the 3 and will be in Arena anyway.

It's actually the gerund form of those verbs that presents the problem! 'Paying' and 'Discarding' work because we dedicated a fair amount of our time for M21 to them, but other '-ing' costs still need to be implemented.

You may cast Scourge of Nel Toth from your graveyard by paying BlackBlack and sacrificing two creatures rather than paying its mana cost.

This allows you to cast it just like it's in your hand. You have that option available as long as you qualify.

BB, Sacrifice two creatures: You may cast ~ from your graveyard

This is an activated ability that you'd have to activate, pay the costs for, then during resolution of that ability, cast ~ for it's normal costs. You'd have to reactivate the ability if you changed your mind and wanted to cast it later. You could also pay the BB, Sac even if your opponent had something like Drannith Magistrate on the board.

So, they're actually quite different abilities.

This is correct. Time to Feed was skipped for tech cost reasons (due to the delayed trigger either being new or being embedded in resolution steps IIRC).

It just comes down to maximizing playing experience in the time we have. If we swap Time to Feed with a similar card, the overall experience isn't changed much, and we get to spend that time on the flagship cards that you really want.

#wotcstaff

Don't assume that implementation issues arise from the game effect in question (in this case, the cast permission)! Scourge was indeed skipped due to tech cost, and I'll give you a hint:

All the cards in historic that use the template "by paying".

#wotcstaff

Many of these "numerical aids" as we call them are ad-hoc. I'll suggest adding this one too. #wotc_staff

Not the WHOLE world. Just my opponent and their stuff!

Never sleep on mono-red, it's the most fun thing to do. :)