# What do I do if mandatory actions will inherently take a very long time to resolve? (Splinter Twin + Tyrant of Discord example)

Basically, if a game is already out of time, and we're within the resolution of a stack, yet the resolution will take an obnoxiously long time, what would the ruling be in a competitive environment where the tournament will have to wait for the resolution? Note that this isn't an infinite effect and thus not a draw, it's not unsportslike conduct - stalling because the game is already at time and that rule is for trying to drag a game to time, and I can pick any number of twins I want, it's my right to use that shortcut to pick an arbitrary number. Don't necessarily focus on deconstructing the example (but feedback on that is welcome), help me out on what the resolution would be.

Here's an example:

Let's say I'm going off with my Splinter Twin combo. I have at least 1 land and a bunch of other permanents. I create a very, very large number of tokens (let's say 10^1000).

My opponent uses Aether Vial to vial in Tyrant of Discord. The resolution of the Tyrant's "enters the battlefield" trigger will take an extremely long amount of time to compute. Consider that we'll be randomly picking copies until a land gets sacrificed, which could mean all my creatures get sacrificed one by one, or possibly none of them do (the land gets randomly picked first).

What are the rules for game time here? Random number generators can't even handle numbers that large, but let's assume we have some way to handle this. I can choose an arbitrarily large amount of Twins to create here. What are the rules for resolving this situation? Even if the game goes to time, 5 turns will not pass, this will take place within one turn's stack, and it's not an infinite loop so the game cannot be declared a draw. Is it acceptable to spend, for example 5 hours, if not several days resolving this effect?

Assume the number is just obnoxiously large enough that it actually takes a computer several hours to spit out a result.

• @Kevin I reviewed the rules and according to my interpretation of the letter of the rules it is not considered unsportsmanlike, I might be wrong though. – Hal T Sep 24 '15 at 20:24
• Trying to rules-laywer the definition of "sportsmanlike conduct" seems like the wrong approach – murgatroid99 Sep 24 '15 at 20:25
• @ikegami They aren't all identical. The outcome will be different depending on when the land eventually gets picked. – GendoIkari Sep 24 '15 at 22:46
• The beauty of this method is that as K gets larger, that probability gets smaller. If K is too big to do useful calculations with, then that probability is so small it's not worth considering. – murgatroid99 Sep 25 '15 at 2:52
• @murgatroid99 even if the chance is infinitesimally small, the number still needs to be computed. The computation time would still be very long. – Hal T Sep 25 '15 at 12:08

If such situation as described by you really applies, only the judge can help in this situation.

I have various examples how a judge handled alike situation, but most of them had one thing in common: It ended in a draw.

The examples that didn't, had in common that the player determined a way to shorten the "choose randomly" process. Here it becomes quite tricky.

In the end the judge with the right authority does have the last word on this and it cannot be generally determined for all tournaments. This is an example of a gap in between the rules that propably needs some fixing for a fair ruling.

There is no rule that covers this situation. So call a judge an he will make a decision that allows the tournament to keep running.

One rule that perhaps could be applied here is that you must be able to represent the game state accurately and clearly. IMO, there is no clear way to represent 10^1000 token. Yes, you could write 10^1000 in a paper, but once a few of those tokens die, then it is no longer clear. If we restrict the number of tokens to a number that can actually be represented in the game, we should be able to manage the resolution of Tyrant's trigger.

• Is it not common practice to call out a large number like "a million twins" when executing an unlimited combo? – GendoIkari Sep 24 '15 at 23:43
• I guess it depends on what you mean by "represent". You could write "1,000,000" to represent a million tokens, and "999,995" if 5 die. But then you could also write "10^1000" to represent that many tokens, and "10^1000-5", if 5 die. – murgatroid99 Sep 25 '15 at 0:14
• Pretty sure it can be accurately represented. – Waterseas Sep 25 '15 at 13:43

The fundamental issue here, is that you and your opponent appear to be unable to identify a mutually agreeable shortcut. As such, it's a slow play violation. I would tend to think the player "at fault" is the player generating the excessive number of tokens. But I'm only the referee in my house, not in tournaments.

None of the short cuts defined in section 4.2 in the tournament rules describe a mechanism for your particular conundrum. But a variety of short cuts exist, each having a different amount of fidelity to the situation.

This particular question about Tyrant of discord has been addressed before. But to summarize. If you have 8 land, and 10^100 other permanents, generate 8 random numbers between [1, 10^100+8]. Those numbers represent the order in which the lands will be destroyed by Tyrant. So, take the smallest of them, and that tells you how many tokens got destroyed.

If you don't have a random number generator available with that kind of range, you can use a 10 sided die to represent the first digit of the number for each of the 8 lands. Then, find the numbers tied for smallest, roll again for the second digit. Etc. Once you get a single number, you can round up, and that will be the number of tokens destroyed.

An alternative approach is to take the average (i.e. mean) number destroyed. That requires a bit of math, but on average, you will destroy (10^100 + 8 + 1) / (8 + 1) tokens, which will leave you with roughly 10^99 tokens. You can drop it down to 10^93 to account for "1 in a million" type chances to placate your opponent, and you should still be good to go.

A final approach, but I haven't worked out the details, is to approximate the number destroyed using a gaussian/normal random variable. You need to compute the standard deviation of the number (it's a computation that's more difficult than the mean referenced above, but can be done.) And then generate an appropriately shifted/scaled normal random variable.

• I think this issue with assigning fault here is that the huge number of tokens by itself isn't a problem; it only becomes a problem once the other player casts a certain thing. The player generating the tokens can't have predicted that the other player was going to do that; and the other player can't be expected to refrain from casting it. – GendoIkari Sep 29 '15 at 14:26
• Your answer explains how the "shorten the choose randomly process" from my answer could be implemented. It could be applied, if the judge actually would agree to it, but it is the judges claim to allow that. There are unforunately no general, always applying rules that could be used in this case. No player actually is necessarily doing a particular violation in here, but this could vary from judge point of view as well. – Aldaris Sep 30 '15 at 7:26