I'm trying to understand why the current HCE policy is crafted like it is. Let me put a not-very-corner-case scenario (happened to me twice).

AP and NAP are two players very aware of the current IPGs, and are playing a mirror of Jund in Modern.

AP controls a Dark Confidant and draws from it without revealing, and draws for the turn too, before NAP notices. No Cheating involved, they call for a judge. This is a GPE-HCE, it's even Example C.

C. A player resolves a Dark Confidant trigger, but forgets to reveal the card before putting it into her hand.

NAP casted a Duress two turns ago and saw cards K1, K2 and K3. Therefore, and supposing nothing was played, AP has another unknown card, U1, and the two cards drawn this turn, U2 and U3.

IPG actually allows us to consider Duress-effects, so U1, U2 and U3 would be revealed, NAP chooses one and Dark Confidant will use this card as the non-revealed.

But it turns that NAP is not sure about the identity of K3. For instance, he doesn't know if it's a Tarmogoyf or an Scavenging Ooze. Or, in a simple example, AP has a Tasigur, Golden Fang in his hand, so NAP chooses not to write down any hand contents, so if AP messes up, he could choose Tasigur for 6 life.

That's only on NAP part. However, it could happen too that NAP has noted cards that AP no longer has in hand (being a mistake writing them, or not crossing them once played, and being now in graveyard or whatever). So, actual contents of AP hand differ from what NAP noted.

Now, NAP thinks he knows three cards in AP's hand, but it's only partially correct. Two are right, one is wrong.

If we remove only two cards from the card set in hand for NAP to choose, he will know that somehow he wrote cards in a wrong way. AP would be angry, since we indirectly corrected a mistake from NAP.

If we remove another “random” card to not give that info to AP, it's deviating largely from the IPGs.

I find this clause of “you can take into account Duress-effects for fixing this infraction” prone to opportunity cheats, and somehow inconsistent with previous policy (we were explicitly said that this info CAN'T be taken into account, to avoid this kind of issues).

Sorry for the wall-text. I'd appreciate any input to help me clarify my thoughts :)

Hello,

I would like just to point out that even if you consider this not to be a corner case, it involves 2 players aware of how HCE are corrected (so probably judges?), HCE actually being done and players are trying to game the policy. That looks like corner to me honestly.

But it turns that NAP is not sure about the identity of K3. For instance, he doesn't know if it's a Tarmogoyf or an Scavenging Ooze. Or, in a simple example, AP has a Tasigur, Golden Fang in his hand, so NAP chooses not to write down any hand contents, so if AP messes up, he could choose Tasigur for 6 life.

Well if he is not sure, he is not sure. It happens, I do not think giving penalty for that should be the norm.
In the case of tasigur, well you can not write down the card and then take to risk to forget it to gain a potential advantage if the opponent does a HCE. Again, that seems very corner and is relying on your opponent making a very specific mistake later.

AP would be angry, since we indirectly corrected a mistake from NAP.

Well, I would say that AP should be angry because he put a card into his hand improperly and was granted an HCE, not because the fix suggested by the IPG can sometimes correct notes from his opponent.

I find this clause of “you can take into account Duress-effects for fixing this infraction” prone to opportunity cheats, and somehow inconsistent with previous policy (we were explicitly said that this info CAN'T be taken into account, to avoid this kind of issues).

Policy does evolve and it has.
I do not know what you call cheats, but not writing down the name of a card is not cheating to me, neither is not remembering a card.
There is very little to be gained here (while again, this edges can be gained by expecting your opponent screwing up and HCEing all over the place) and being inconsistent with previous policy is obviously the case, since it has changed.

This new method corrects the absurd cases with 3 creatures in hand after the HCE made by not revealing on Xenagos or Ajani, of 4 nonland spells in hands after doing a +1 on Origins Nissa.

This policy has been crafted to correct game play errors the most fairly possible in most cases. Knowing the IPG has always granted edges to players with that knowledge but I don't think that it is world breaking. After all, it has been said again and again, but this is better than Game losses that used to be granted for the same infractions. I guess that if you want to find cases that breaks the policy, you can, but the cases you exposed do not really worry me.

I think AIPG cover it perfectly:

Originally posted by IGP:

Information about cards previously known by the opponent, such as cards previously revealed while on the top of the deck or by a previous look at the hand, may be taken into account while determining the set of cards to which the remedy applies.

AIPG
Even though the contents of a previously revealed hand are not public information, the knowledge of those cards, when agreed upon by both players (such as from notes taken about the hand) may be useful in reconstructing as correct a game state as possible. There is a “may” here, however. If there is doubt or disagreement about what has been revealed some turns back, and you aren’t confident that the set is correct, you may treat the entire set as the grouping to which to apply the remedy.

Edited Bartłomiej Wieszok (April 20, 2016 09:18:32 PM)

Ok, I now understand that I'm probably worrying too much about this, and in most cases, it provides a better solution for the problem. Thank you!! :)