20FEB18 - Defy Death Update

Nope, sorry.  

The Defy Death skill now does what it said it would do, rather than what it accidentally did back in the day :)

Best,

SC

An update/patch to the commander Defy Death skill (Military skill tree).

We've recently received reports that the commander skill Defy Death works erratically, and/or over/under-enthusiastically, in bursts.

Commanders with a Level 10 Defy Death should only survive 10% of the time, however (anecdotally) we've been told that in most battles, when one commander belonging to a player army dies, most all of them die; when one commander survives in the same player army, most all of them survive.

We looked at the historical logs, and this anecdote proves true.

This was a code issue that has been in place since commanders were introduced to the game back in 2010 and has affected all players, attackers or defenders, with equal likelihood, in every combat.

Having delved into the codentrails, and after extensive consultation with the Norse God of Mischief, I've determined that the error was in the RNG (Random Number Generator), which wasn't providing as random a number as we would have expected.  If that's not oxymoronic.

Therefore, I've rewritten the RNG and deployed it to live, and it is now truly random - or at least as random as any logic-based computorial system can be.  Please feel free to reply with any ramblings you may have about pseudo-RNGs, chaos theory, ordered systems and/or the Second Law of Thermodynamics below.

The (few) other systems that depend on RNG (such as diplomatic chances, theft etc) were unaffected by this bug; and continue as before.

Also, during the investigation, it appears that commanders were *much* more likely to survive, with a likelihood getting close to 50% in certain circumstances.  Of course, the stated aim of Defy Death is a 1% chance per level, up to a maximum of 10% at Level 10, and this is now in place and working as it should; so please do expect a more realistic 1-in-10 chance of surviving at L10 Defy Death.

We're actively monitoring the new RNG and the diceroll outcomes it provides at a statistically significant level, and it appears to be working as it should, now. As things stand (the new RNG has been running for a few days) we have data from 11,817 commander Defy Death rolls, and 1,138 of them are <= 10%, which seems statistically significantly likely that things are now working as intended.

tldr; The random number generator for deciding whether military commanders should Defy Death or not was a bit broken, and has been weird ever since commanders were introduced to the game.  This has now been fixed, and commanders should now Defy Death in keeping with the expected 1% per level of the Defy Death skill, at an individual level of granularity.

Best wishes,

SC