Dumb Question #22194

[Tamur]

hi all, I should know this by now, but here's a situation:

If you've modded unit hit points and terrain movement speed, do a custom battle, save the replay, and send it to someone who has a different mod of the unit and movement files, will they see the same replay you're seeing?

The more basic question is, are replays based on detailed recorded information, or are they based on a re-generation of the battle based on a simple recording of commands by the AI and the player?

[Kekvit Irae]

I have never played recorded replays of others in RTW, but I can guess from my experience in MTW that the answer would that the person using an unmodded version of the game (or a differently modded version) would see a completely different version of the replay than you.

In MTW, I once viewed a replay about French Knights charging billmen and other spear units. They were "supposed" to be wiped out, but I was running my KekvitMod which adds charge attack to French Knights. The cavalry cleaned up the units and then acted wierd.

[Tamur]

Thanks. This is what I was guessing, having seen the same behaviour with MTW as you describe. Guess I could try it with the -mod switch

[Captain Fishpants]

AFAIK replay files don't save data directly relating to unit stats, so if you create a replay, then mod the unit data, then play the replay things will get very odd very quickly. The replay will soon "get out of step" if the unit data has changed, because, for example, all the combat calculations are being done while the replay is in progress. Modded units won't produce the same results.

Providing that nothing changes between recording and replay everything should be OK, as you'd expect.

[Tamur]

Thanks CF, had not thought of just doing a temporary mod of these files to test.

Another question on this subject for anyone: are the two files...

descr_battle_map_movement_modifiers.txt
export_descr_unit.txt

...the only ones used by the replay engine?

edit: I guess terrain would be in there somewhere, but the likelihood of terrain being modded is fairly near nill.

[Papewaio]

AFAIK replay files don't save data directly relating to unit stats, so if you create a replay, then mod the unit data, then play the replay things will get very odd very quickly. The replay will soon "get out of step" if the unit data has changed, because, for example, all the combat calculations are being done while the replay is in progress. Modded units won't produce the same results.

Providing that nothing changes between recording and replay everything should be OK, as you'd expect.

Does this mean the replay has the 'rolls' used in calcs not just the results of the rolls?

[Tamur]

Good question! I could guess but that would be fairly worthless.

[x-dANGEr]

Hmmm.. That simply justifies the lag in the replay if the battle was laggy..

Example: Put GFX on max, men on hugest (Through data files), all peasants VS all peasants. The battle will lag, save the reply and watch it, it will still lag.

[Severous]

A bit off topic but an observation about replays...

I recorded a battle and when replayed there was a different outcome. All seemed to hinge on death of the general which happened in the replay but not the original recorded battle. I hadnt modded anything between.

[Tamur]

Captain Fishpants' reply seems to indicate that this difference should only happen when the unit or terrain files are modified between the save of the battle, and the replay.

Different outcomes happened fairly constantly in 1.0/1.1 for me. So far I haven't seen the same problem with 1.5, but then again I've only done about ten tests cases rather than dozens or hundreds.