Classic Buzzing Crash + in the Dark = Unhappy Player

So there I was, strolling through the dark in a city node (I feel confident enough after the last patch that the stairs will be open) and yikes, I get the dreaded screen freeze, 5 second buzz, and then crash out of POE. And upon re-entering the game - yup, the 10% Death Penalty had been applied (the unhappy part - the rest is just Standard Operation Procedure by now). Since I was in no distress prior to the freeze, I am assuming that the server treated me as just standing there in the dark and had enough time to kill me.

My point: from a player's point of view, with (seemingly) random crashes occurring rather too frequently in the Delve, something in how the server handles death on crash really should give (even if it just a temporary switch that gets reset in "normal" times).

Please note that I am reporting this as a design issue in the suggestions forum and I am not making a bug report. I have no wish to be hassled over perfect form regarding bug reports.

Haha the darkness before your eyes when you loose 10% because of the games fault 😃

Haha the darkness before your eyes when you loose 10% because of the games fault 😃

I'm not sure how the game handles the processes of a crash, since crashing is kind of an exception to any written rules obviously.

However what tends to happen during crashes is the same thing that tends to happen when you Start+Y to avoid a death, packet loss. Two seperate issues im sure, but the effect is the same.

As a crash is an exception to written rules, I'm not sure how much GGG could design around it.

"

Avic_Hybana wrote:

I'm not sure how the game handles the processes of a crash, since crashing is kind of an exception to any written rules obviously.

However what tends to happen during crashes is the same thing that tends to happen when you Start+Y to avoid a death, packet loss. Two seperate issues im sure, but the effect is the same.

As a crash is an exception to written rules, I'm not sure how much GGG could design around it.

You are quite right of course - no one but GGG (and perhaps some ex-employees I guess) really knows how their software handles a client side crash. In the case where the game instance ends prematurely but the client stays active, perhaps there is nothing to be done. But GGG does seem to know when the entire client crashes (they auto report it) and in these cases they would also know at what point communication was last received from the client. Perhaps in the case of a client crash, that time of last communication could be used as the time for the restore point?

Just an observation from various leagues I've played where I have experienced crashes, which is every league, just about every day, sometimes multiple times a day...

The game reporting of direct player-account interactions with dynamically generated "items" is solid. IOW, the item that I picked up just before a crash will be in my inventory. I have never experienced a crash where I feared that an item I picked up would not be in my current character's inventory. ie: After a crash, I have never experienced a "roll back" to a previous state in terms of player items/inventory and/or things directly linked with my "account."

Clicking on chests/interactive components appears to require some interaction with something outside of the client. ie: If you click on a chest and nothing happens, don't keep clicking on it! Somewhere, some demon is very angry you clicked on that chest and is being slow about doing anything about it... unless you keep clicking. At that point, the demon will really get angry and crash your client. Why? I dunno. Maybe every single click on that item requires a new communication with something and repeated clicks just make it angry? It may mean that dynamic generation of chest contents require server interaction and continuing to force those interactions when a glitch has occurred is A Bad Thing To Do ™.

Sometimes I'll loose access to an Atlas map I was in when the crash occurred and sometimes I won't. What does that discrepancy mean? It's obviously a different response with the same gameplay variable, so why is that? Something "different" is going on there and either not all crashes are the same or not all crashes are handled in the same way. Why is that and how can we use that info to help keep our games stable?