The meaning of numbers

[Monkwarrior]

Hi
I am polishing now the first version of the Reconquista Mod and I am trying to regulate the type of units that will appear in the different rebellions.
These data appear in the column 11 (from Gnome editor) and it assigns a number for each type of rebellion available for a given unit. For example
"CATH_REBELS(15), CATH_LOYALISTS(5), CATH_PEASANTS(10), CATH_BANDITS(12), CATH_ZEALOTS(10), ORTH_REBELS(15), ORTH_LOYALISTS(5), ORTH_PEASANTS(10), ORTH_BANDITS(12), ORTH_ZEALOTS(10), ENG_BANDITS(2), CATH_HERETICS(15), ORTH_HERETICS(15), PAG_REBELS(10),

Can anybody explain to me the meaning of those numbers? I guess they are related with the probability for the unit to appear in a rebellion, but I don't know if the probability increases with higher numbers or with low ones and if these numbers have limits.

Something similar happens in column 15 about AI choices. For example
"POVERTY_STRICKEN(256.5), DESPERATE_DEFENCE(17.1), CATHOLIC_EXPANSIONIST(136.8), CATHOLIC_NAVAL_EXPANSIONIST(136.8), CATHOLIC_TRADER(171),
I guess these numbers are related with the probability for the AI to produce this type of unit in case of different behaviours of each faction, but now the numbers are much higher than before.
Any idea?

[[DnC]]

1. The numbers are merely a value of how many of that particular unit spawns in a certain rebellion/crusade/jihad/appearing mongols etc etc. The more negative a number is, the more of that unit spawns in whatever it spawns in. At least this is my experience. Then again I've only actually tested it with the MONG_RAIDERS and it seems to work like that there.
So the more negative (e.g. -10), the more it spawns and the more positive (e.g. 10), the less it spawns. Although I could be wrong.

2. The higher the value, the more likely the AI will build it, but there are other factors involved of which I know at least one; the buildcosts. There are more factors I think, but can't quite think of them at the moment.

Have you read about Duke John's bug discovery relating this? If not, click here.