Discussing Monks

Age of Empires II: DE II - Discussion
22

I recently had an Idea how monks could be made free of rng while still keeping the “one by one” conversion effect.

The core of this idea are 3 added ressouces to the civs.
The idea is only for converting units, buildings should have their own system.

The most important change is that every unit will get a conversion threshold. This is how many monk seconds a single monk would need to convert that unit.

One that tracks all currently active conversions. Just the sum of them.
Let’s call that n.

One that tracks the accumulated sum of all currently invested “monk juice” into conversion. So every time a “monk second” is applied to a conversion, this ressource goes up by 1. If a conversion is successfull the needed threshold for that conversion (the conversion threshold of the converted unit) is substracted from that ressource.
Let’s call this j.

Lastly one ressource that counts the sum of all the ressource threshold of the currently targeted enemy units to convert. Whenever you target an enemy unit with one of your monks that unit conversion threshold is added and whenever the conversion is interrupted or the unit is successfully converted the threshold is substracted. (Optimally it just adds up all currently active conversion thresholds togehter as techs like faith or teutons bonus can change the threshold while a conversion is ongoing)
Let’s call this t.

So how the game determines if a conversion is successfull?
When a Monk “shots” its Monk second on an enemey unit, first 1 is added to j.

Then there is the proof if j >= t * (n+1) / (2n). If this is true, the conversion is successfull, otherwise not.

This is a compromise between “A lot of conversions lead to a lot of “lucky” ones”" and “A lot of conversions are all “unlucky””. If there are a lot of conversions ongoing this will result in the first conversion being basically twice as fast as a 1 by 1 conversion, but once there is only 1 conversion ongoing, this one will basically need another full conversion period to happen. I have 2 examples for 5 or 10 units with a conversion threshold of 10 seconds, how long it will take for each conversion to happen.

Conversion 1 2 3 4 5
at Second 6 8 9 12 17
Conversion 1 2 3 4 5 6 7 8 9 10
at Second 6 7 7 8 9 10 11 13 15 20

On averagethe convervsion time is still 10 seconds (if you don’t account for rounding always up due to the “monk second” mechanic. (Achieving the same average as for single conversions is the result of that formula I postponed at the beginning, t * (n+1) / 2n. This inverses the effect of the staggered conversions over one single ressource.

The effect is for high amount of conversion similar to the randomness of the current conversion mechanic. But with the difference that there is absolutely no randomness involved.

Ofc this will also have some side effects I haven’t looked into much yet.
One is if we have conversions on units with different conversion thresholds. Converting a unit with a high conversion thresold will increase the conversion time of ALL units until this unit is converted. But also the other way around, if you try to convert units with high conversion resistance you can also target some other enemy units with a lower resistance, this will reduce all conversion times. There is then also no preference, so if you target a unit with a conversion threshold of 5 and one with 10, after 6 seconds you will convert either of them, completely neglecting the individual conversion thresholds.

I think this can be branded as a “feature”, open to figure out for the community if there are some advantageous strategies in your targeting.

Edit: Note that there is stil no randomness involved. Which unit will be converted will be determined by the order and timing in which you target the enemy units but aswell the movement of the units, when they are in conversion distance and when not. It’s just that there is no preference, not that there would be no way to influence. It’s only not clear to me if there is an advantageous strat like "target first 5 “hard to convert” units and then 5 “easy to convert” units or something like this to get faster conversions on hard to convert units or whatever.

Then there is the question what happens with the “loaded up” j ressource if the opponent completely disengages and there are no ongoing conversions. First, when this happens it is very likely that you will get some conversions of the disengaging enemy. Cause you will load up your j value. When then the number of ongoing conversions is reduced, the t value goes down while you still have a high j value. Resulting in a few conversions of the disengaging enemy forces. But what happens if for some reason (maybe monks killed or whatever) the j value is still high but no conversion ongoing?
I would say, that if n = 0, then the j value if it’s bigger than a set maximum value (for example 3) will be tuned down to that max value. I think this is important, so even IF this happens, there will still be no insta conversion possible, just a faster conversion than regular (faster by that amount of monk seconds).

For buildings I would make a completely independent conversion system. Also with an indicator bar how far the current conversion is actually, so you know when it makes sense to delete the building and when you can try to kill the converting monk. I don’t think it’s good to “charge” your conversion on buildings before switching to units.

Edit: This method would also allow for much more sophisticated unit designs, as it would allow to give each unit a completely independent conversion threshold. So slower melee units could eg. get a way higher threshold than faster ranged units to compensate for their harder / easier matchup vs monks. Also more expensive units like eles could just get a way higher threshold so monks would just need a bit longer to convert them.

23

I really like this idea. I think it could be even simpler though. Just tie the conversion and faith regeneration speed to the cost of the unit the monk is converting. No random luck involved. Cheap trash units convert really fast and the monks regenerate fast. More expensive gold units like Elephants take longer to convert and longer to recharge faith. All units have a resource value cost which directly correlates in linear fashion to a corresponding conversion time.

1 Like
24

I think this example is wrong, or I didn’t fully understand your proposal yet. Let’s consider the first example with 5 units with a conversion. That means there are currently 5 monks trying to convert? Doesn’t that mean, that the monk juice counter j is increased by 5 each second? As such, it would need only 2 seconds to get the first conversion?

I think your idea is originally designed for 1vs1. I haven’t made any thoughts yet on what this system would imply for 2vs2, etc.

1 Like
25

I tought a similar approach, but I tie the conversion resistance with the gold cost of the unit in an inverse way. As the unit cost more gold then the less convertion resistance it have

Trash units are more resistance, so, counter monks. I think trash units should be more loyal than “paid” or professional units (gold units).

1 Like
26

The issue with that is that it doesn’t allows for “fast” conversions (unless you use multiple monks. Then Theocracy can become absurdly overpowered for once also).
Before that tech monnks would be very awkward with no “fast conversions” possible or reducing the amount of possible conversions heavily. The opponent could repeatedly threaten an engagement until he gets a favorable one with the monk player having to react all the time very sharply.
Vice versa once Theocracy is in, monks could become absurdly overpowered as you don’t even get fast conversions but can make them in rapid succession.

So what I tried to accomplish here is some kind of “compromise”. With only a single monk converting a single enemy unit it is a set time until the convesion happens, but with higher numbersyou will get some fast conversions like you get currently. It’s just so that it isn’t random anymore. You don’t get one time 4 random fast conversions or none but instead it’s only dependent on your and your opponents skill.

I expected this question somehow. I try to make the math a bit cleared:
Wuth 5 ongoing conversions against 5 units with 10 conversion threshold each you get:
t = 5 * 10 = 50
n = 5
Which means for the first conversion the j value needs to be
j >= t * (n+1)/2n = 50 * 6 / 10 = 30
With the 5 monks each adding 1 to the juice ressource every second, it needs then 6 seconds for the 5 monks to get to the j = 30 needed for the first conversion

After that:
t = 40
n = 4
j = 30-10 = 20
Then j needs to be bigger or equal to 40 * 5 / 8 = 25. With 4 converting monks this will take 2 more seconds as 20 + 4 is 24.

To summarize, the first conversion in this scenario is applied when you have collectively invested 30 monk seconds in the conversions. Then the second after 25 monk seconds (but note that 10 monk seconds are substracted from your ressource once you have converted the first unit) and so on.
The Result of that calculation is the depicted times for the conversions to happen. There are no basically instant conversions but instead a similar behaviour like currently. For some time the converted units are “safe”, but once that safe time is over there will be some conversions in rapid succession. Then the conversion rate will slow down a bit.

Idk if it needs to be that strict. I think for the sake of Balance in the lategame it would actually be better if only scouts had a good matchup vs monks of the trash unit. Otherwise monks will fall down very rapidly in the lategame.
We already have that effect Look at Bohemians and their “Hussite Reforms” tech. Despite making Monks trash and spammable like crazy in the very lategame basically nobody uses it. And it makes sense as Monks are already quite bad vs trash. Even against halbs and skirms. Skirms can kill lategame monks cause no armor, but both units are also very easy to replace for the opponent. And then we also have hussars that slaughter monks.
So monks don’t provide much in the very lategame, they are fairly easy to counter with a unit you want to make anyways to raid and then the micro is also a nightmare at this stage of the game.
So I don’t think it’s a good idea to make monks even worse then. I just think it can be probably benefitial to make some units more resistant to conversions than others. Best example is ofc the elephants.

Ofc it’s designed for 1v1s. But I don’t see how it would make a big difference in multiplayer. Not to say that monks are not necessarily a good multiplayer unit anyways cause… yeah archers counter monks quite well.

1 Like
27

Not all trash units should counter monks. Just scouts. Otherwise it ruins the game balance.

1 Like
28

Thank you, I see my error now.

I want to show you another example. To avoid decimal numbers I use an equivalent conversion condition. Instead of
j >= t * (n+1) / (2n)
I use
2 n j >= t * (n+1).

Let j=0 and n=5. Scenario:

  • Theocracy is researched, i.e. if a group of monks convert a unit, then only one monk needs to reset.
  • The 5 monks target focus the same unit.
  • After successful conversion the remaining 4 monks move on and target focus another unit and so on.
  • Suppose the target units have all a threshold of 10
  • Suppose there is no delay introduced by the player (time required for necessary commands etc).

Then I get these values, if I’m not wrong:

Second n j t 2nj t * (n+1) conversion
1 5 5 10 50 60 no
2 5 10 10 100 60 yes
2-3 four 0 0 - - -
3 4 4 10 32 50 no
4 4 8 10 64 50 yes
4-5 3 -2 0 - - -
5 3 1 10 6 40 no
6 3 4 10 24 40 no
7 3 7 10 54 40 yes
7-8 2 -3 0 - - -
8 2 -1 10 -4 30 no
9 2 1 10 4 30 no
10 2 3 10 12 30 no
11 2 5 10 20 30 no
12 2 7 10 28 30 no
13 2 9 10 36 30 yes
13-14 1 -1 0 - - -
14 1 0 10 0 20 no
15 1 1 10 2 20 no
16 1 2 10 4 20 no
24 1 10 10 20 20 yes
24-25 0 0 0 - - -

Two things I want to highlight:

  • at time “4 seconds” there is only 8 monk juice accumulated, but the conversion is successful
  • in my example there are sometime negative values for j. I’m not sure if this is intended or if j should be capped at 0.

Thus, in this example the conversions take place at time instants:

Conversion 1 2 3 4 5
at Second 2 4 7 13 24

Comparing this to the 1 monk - 1 target example, which yields:

Conversion 1 2 3 4 5
at Second 6 8 9 12 17

Thus, at first the target focussing converts the first few units faster but ultimatley it takes longer to convert all 5 units (intersting observation, not a criticism)

31

Maybe I used the wrong word. Trash would be only more resistance to conversion, but not would have bonus against monks.
I don’t think balance will be strongly affected, because monks mainly use is against knights, elephants and some no trash UUs

34

What if when a monk starts converting, it slows down the unit being converted? Or would that be too OP?

35

Another interesting scenario that came to my mind:

Let n0, j0, t0 be the current values of n, j and t, respectively. Let them be such that
2 n0 j0 < t0 * (n0 + 1)
holds.

Scenario A:
At the next second the conversion condition is satisfied if
2 n0 (j0 + n0) >= t0 * (n0 + 1)
holds.
This can be reformulated to:
2 n0 (j0 + n0) - t0 n0 >= t0

Scenario B:
However, suppose a monk got killed in the meantime, i.e. there are only n0-1 monks left. That means, for a successful converstion this condition must be met instead:
2 (n0 - 1) (j0 + n0 - 1) >= t0 * (n0 - 1 + 1),
which is equivalent to
2 n0 (j0 + n0 - 1) >= t0 * n0 + 2 (j0 + n0 - 1)
and can be reformulated as:
2 n0 (j0 + n0) - t0 n0 >= 2 (j0 + n0 - 1) + 2 n0

Question, to which I have not an answer yet:
Is it possible that in scenario B a converstion takes place but the conversion condition is not yet met in scenario A? Or put in other words: Is it possible that the killing of a monk results in an earlier conversion?
Mathematically formulated: Are there (n0, j0, t0) such that:

  • 2 n0 j0 < t0 * (n0 + 1)
  • 2 n0 (j0 + n0) - t0 n0 < t0
  • 2 n0 (j0 + n0) - t0 n0 >= 2 (j0 + n0 - 1) + 2 n0
36

It’s not a question.
Given that all units that are in process of conversion have the same conversion threshold.

When a monk is killed in this process and j is bigger than half what is needed for a conversion this will result in a faster conversion of the next enemy unit.
The same is valid for every form of interrupion to the conversion. That menas, whenever the opponent disengages it is very likely that a few of the units will be converted in the process, depending how long the conversion was already ongoing. If the opponent disentgages well, so that all conversions will be interrupted at the same time, there will be no conversions, but in most cases a disengaging opponent will lose some units. Ofc with this basically guaranteed behaviour it’s clear that the average conversion time needs to be increased to compensate this.
It would be a new “feature” for the monks, that every time you successfully target multiple conversions you are basically guaranteed to get a conversion if the opponent doesn’t disengages immediately.
But it is also fair, cause it benefits very fast reaction times of both players. Atm when your units are converted by the opponent you have about 2-3 seconds to react, with this new behaviour if you react only after 3 seconds you are very likely to lose a few of your units. But if you manage to react basically immediately you won’t lose any. So there is a more graduate payoff system for your actions.
It’s only important to check that there istill has to be a benefit in disengaging in most cases and be it only by buying some time to get other military on the field so you can deal better with the monks.

The other thing that I checked was if the units have highly different conversion thresholds: If one unit with a very low threshold is converted that then the next one could be also immediately converted. But this is seemingly not the case. I need to make a proof for that though. But it looks that every additional conversion still requires some more monk juice to be invested, what is exactly how it should be in my opinion. It would be weird if the leftover juice after a successfull conversion would be higher than the needed juice for the next one 11.

One issue that I haven’t discussed is actually Theocracy. With the described calculation it would still be very strong (though because of the calc method the conversion threasholds can be chosen a bit higher than it would be equivalent to the current conversions) - but with no limit to the minimal conversion time theocracy is still quite powerful. ofc you can argue that with multi-targeting you will get at least the same amount of conversions when the opponent disengages.
One way to reduce that theocracy effect a bit would be that when converting with multiple monks, only the first monk gives a full monk second and all further monks only half a monk second. Like it is with vills when you rush up buildings.

37

Monk RNG is a red herring in terms of the chance that it actually changes the outcome of a game. Its likely a tiny number of game outcomes changed.

Relative to a deterministic counterfactual the expected outcome will remain the same. Which means for almost every interaction the decision for the opponent to commit to killing the monks or delete or run away or whatnot remains unchanged (assuming risk-indifference). Even if you explicitly analyze the cases its clear monk RNG isnt that different from other forms of RNG in the game.

If its an expected conversion:

  • non determinstic: the smart money says delete or run away or sacrifice some cheap unit to conversion. Its not like its smart to try and gain an upset here vs the monks. Monks might get a conversion if chosing to run away.
  • deterministic: same outcome. Delete, run away or sacrifice. Upsets are basically impossible and monks cant get conversions on running away units.

If its not an expected conversion:

  • non-deterministic: possible monk upset (see below). Lose some units based on CDF for monks.
  • deterministic: dive + kill the monks. Guaranteed win.

So the only real way the change would manifest is in making monk upsets less likely. But since this only occurs in situations where there is not an expected possibility of conversion it is entirely the prerogative of the non-monk user to engage or not. Which means the choice to engage in the risky activity is completely avoidable.

The risky behavior of diving against a monk and expecting a kill is also down in the noise relative to other RNG things. First and foremost is map gen. Engaging in mangonel fights is realistically statistical in nature and either side can choose to disengage. Even in larger scale fights, if the battle is close it will often be random who wins which is remarkably similar to how engaging monks works. Choosing to engage monks when you dont have a significant advantage is like choosing to fight an enemy army which is very close in strength to your own. Theres going to be upsets that occur from time to time due to things like a lucky scorpion bolt. Heck AoE2 is an imperfect information game and lots of times you win or lose based on the luck of decisions made based on the statistical nature of how players behave. This manifests itself as upsets in terms of unexpected strategies from the opponent. But thats just the nature of mixed strategies.

In fact the very nature of monks means the probability of large-enough-to-affect-the-game upsets diminishes rapidly with increasing numbers. The games where monk RNG is so lucky that even with a comfortable margin when diving vs monks the game gets completely turned around is extremely rare. I cant even remember such a game ever occuring. I remember a lot of times where a player chose to take the risk for an expected kill vs like 2 monks and lost. But I cant remember any game where this wasnt just down in the noise by the end of the game.