Heroes of Might and Magic Community
visiting hero! Register | Today's Posts | Games | Search! | FAQ/Rules | AvatarList | MemberList | Profile


Age of Heroes Headlines:  
5 Oct 2016: Heroes VII development comes to an end.. - read more
6 Aug 2016: Troubled Heroes VII Expansion Release - read more
26 Apr 2016: Heroes VII XPack - Trial by Fire - Coming out in June! - read more
17 Apr 2016: Global Alternative Creatures MOD for H7 after 1.8 Patch! - read more
7 Mar 2016: Romero launches a Piano Sonata Album Kickstarter! - read more
19 Feb 2016: Heroes 5.5 RC6, Heroes VII patch 1.7 are out! - read more
13 Jan 2016: Horn of the Abyss 1.4 Available for Download! - read more
17 Dec 2015: Heroes 5.5 update, 1.6 out for H7 - read more
23 Nov 2015: H7 1.4 & 1.5 patches Released - read more
31 Oct 2015: First H7 patches are out, End of DoC development - read more
5 Oct 2016: Heroes VII development comes to an end.. - read more
[X] Remove Ads
LOGIN:     Username:     Password:         [ Register ]
HOMM1: info forum | HOMM2: info forum | HOMM3: info mods forum | HOMM4: info CTG forum | HOMM5: info mods forum | MMH6: wiki forum | MMH7: wiki forum
Heroes Community > Library of Enlightenment > Thread: Tribute to Strategists
Thread: Tribute to Strategists This thread is 20 pages long: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 · «PREV / NEXT»
Laser70
Laser70


Adventuring Hero
posted October 10, 2015 05:23 AM
Edited by Laser70 at 06:11, 10 Oct 2015.

You outlined the details much better than I did, just about 100% of the forum are studying the numbers around the game as if they already understood why they shouldn't. Most people know, but don't grasp (big difference, biggie)

But it doesn't have to be a rule, if players are new to the game, by all means study it, the game has a limited replay value anyway. To me personally, the game rapidly went downwards after only a few months of studying the mechanics, that was a huge mistake. Now I have to find a new game to play because I never lose a battle, actually I do lose battles, but it doesn't feel like that anymore, which means I almost never lose battles.

I could always increase the quantity of AI towns and thereby try to compensate with more enemies rather than more quality in the enemy, but I absolutely hate massive battles, I love the tiny battles, especially early games, I love walking around with very tiny unit stacks and low hero stats. As soon as I reach over 500 units in a stack, I quit the game and create a new map.

One thing that would do wonders to this game is if someone would create a new AI system with a learning AI with true ability to learn from its mistakes.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted October 10, 2015 06:10 AM

Quote:
I agree, when I found out that Santa didn't exist it totally ruinned Christmas.

It's all about your own inner rules you lean upon when deciding which things are OK to believe in. I presume, you've never been to outer space, yet you likely believe things you're told about it. I would rather call it a blind faith, as you blindly accept knowledge from 'authorities', being unable to check whether they lie or not for yourself. We all have a faith of our own.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
Laser70
Laser70


Adventuring Hero
posted October 10, 2015 06:13 AM
Edited by Laser70 at 06:58, 10 Oct 2015.

Instinct is a good thing, then there are entire books about instinct which bring in a thousand different new perspectives. Idiots know why they shouldn't hit someone in the face, and in 80% of the cases they won't hit someone in the face, as soon as they read a book about it that percentage increases up to 99%, if not 100%. People 'feel' things, but don't know how to explain it.

It's a little bit like that in the gaming world too, people 'feel' they shouldn't do certain things in a game, and yet, you find evidence that most people do them still, until they understand why they shouldn't.

A good example is if you're being approached by a moose, you know instinctively why you shouldn't pick a fight with it, a child knows it too, but if you study the moose you get a pretty darn clear picture why all the why's.

The guy who have suspicion will obey that suspicion partially, the guy who knows will form strong principles. The people on the forum have no principles, they study the numbers, so they are driven by suspicion they suspect it might ruin the game but aren't quite knowledgeable to say 'no'.

Some people take risks (What the hell, I'll study them anyway, and continue to see whats going to happen I don't care) and then they continue, not quite sure whether or not the game will be ruined or not. And then you have a very few people who are absolutely sure that the game will be ruined, they posess knowledge that it will be ruined.

I wouldn't say that it is even about knowledge, because many players have studied the game more than I have done. I think there is more to it than just knowledge, you have to have something which acts on that knowledge in a more effective way because many people have studied the game for 15 years, I studied it for 3 months. Our brains are strange things, some people are still entertained by something after 15 years because of the simplicity that is needed to entertain it, other people get tired of it after only a very limited amount of time.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
bloodsucker
bloodsucker


Legendary Hero
posted October 15, 2015 05:10 PM

AlexSpl said:
Quote:
I agree, when I found out that Santa didn't exist it totally ruinned Christmas.

It's all about your own inner rules you lean upon when deciding which things are OK to believe in. I presume, you've never been to outer space, yet you likely believe things you're told about it. I would rather call it a blind faith, as you blindly accept knowledge from 'authorities', being unable to check whether they lie or not for yourself. We all have a faith of our own.

Man, or you're way more subtle then I am able to deciver or you completly lost my point.
I meant, intuitively or sistematically you can't help yourself to learn some mechanics of a game you play and then have to take the inner decision of continue enjoying while playing on a more resourcefull way or quit. There are no other choices.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted October 15, 2015 05:38 PM

Knowledge about game mechanics can't hurt you in any way if you a reasonable player with no lack of imagination, it can only sharpen and enrich your experience if used thoughtfully and if you're not seeking knowledge just for knowledge itself.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
batoonike
batoonike


Known Hero
posted November 07, 2015 02:38 PM

Laser70 it's possible for the game to become better the more you know about it. Good games do that, if the player is also creative. Sure it get's very easy and you can beat 7 AI-s in a team on "impossible" diffuclty without even thinking but you can always create a bigger challenge. How about 7 AI-s in a team and you are not allowed to buy units or use artifacts or spells? It becomes a different game but it's always possible to create a bigger challenge than the previous one and then start solving it.


But I agree that learning stuff generally ruins the experience to some degree. For instance I played minecraft and dont starve. In minecraft ja was always reading the wiki so I knew most things after like a few months. In don't starve I refused to read the wiki except maybe very rarely and I still keep discovering occasional stuff after years.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
RoseKavalier
RoseKavalier


Admirable
Supreme Hero
posted January 31, 2017 06:44 AM bonus applied by angelito on 11 Jun 2017.
Edited by RoseKavalier at 00:10, 03 Feb 2017.

On AI stacks splitting

Inspired by my own Let's Plays (more specifically this post), I recently decided to try and figure out what was the cause of AI splitting its shooter troops (or not).

Let's look at a few more examples...


This doesn't seem to make any sense. Here is a brief list comparing # of Champions to # of AI stacks from the same conditions.
1-4
2-2
3-1
4-1
5-1
6-1
7-1
8-7
9-3
10-2
...

This behaviour is nothing like what we normally expect from Adventure Map creatures - there is a clear comparison of army value to Wandering Creature value and then the split is done (see Tribute to Strategists).

In this post, I will try to explain when and how AI splits its shooters (I haven't looked in the non-shooter cases yet).

Special thanks
Gu7979gu at the Gamerhome forums, this post is pretty much a translation of his findings over here. Chinese to English is difficult but with translating tool I finally was able to make sense of it.

Proper conditions
First, we must know when an AI is willing to split/combine troops together, there are 2 such cases:
1-the AI hero has higher Tactics skill level than you
2-the AI hero attacks you (regardless of your and its Tactics skill, Cheater!!)

Examples of each case:
1- you have Basic Tactics, AI has Advanced Tactics
2- you have Expert Tactics, AI has no Tactics

TL;DR If you are a mapmaker having difficulty with an AI hero combining troops when he should not, then try to make certain neither of these 2 scenarios may occur. A simple solution is to put the hero on Stand Still and not give it Tactics. You can also modify the way AI splits its troops by changing the number of creatures in your strongest stack or by modifying your attack and defense primary skills.

To split, or not to split
Beware, there be math ahead!

What is required:
[*] AI value of creatures
You can find this in many places, notably in this Tribute to Strategists.
[*] A Calculator and some patience (suggested: spreadsheets for convenience)

First step: Calculate the power correction factor.
For this, we need to know the human and AI's Attack Skill and Defense Skill. This is fairly easy, they are shown on the hero mini-tab.


Then we obtain the power correction factor as follows:


Second step: Calculate AI value of your strongest stack
Calculate the value of each stack, and pick the highest. This is basically (AI value)*(number of creatures in stack). Don't count separate stacks together, only single stacks.

We'll call this value HUMAN.

Third step: Calculate AI value of AI's strongest shooters stack. Do this the same way as in second step.

We'll call this value AI_SHOOTER.

Fourth step: Calculate the corrected HUMAN factor.

Use step 1 and step 2.

This step doesn't seem to make any sense (and it doesn't, appears to be a 3DO bug)

Fifth step: what if AI is very strong?
You may wonder what happens when AI_SHOOTER > 65,535... since the HUMAN_C factor is capped at 65,535, we also need a magnitude modifier 'm' to compare apples to oranges.


This modifier will be applied so that the order of magnitude of HUMAN_C and AI_SHOOTER factors coincide.

Sixth step: can I split yet?
Compare steps 3 and step 4 with the 'm' factor from Step 5:


This is pretty explicit. Notice that the magnitude correction allows us to compare HUMAN_C and AI_SHOOTER even if AI_SHOOTER is much larger than 65,535.

Seventh step: ok, how many stacks?
If step 6 concluded that the AI can split, we now decide how many stacks will be generated. Otherwise, the AI remains with 1 stack.

This step should seem more normal... calculate the ratio of step 4 and step 3 and this will tell you how many stacks will be generated.


There's an easier way to remember this:


Notes
Step 6 is most certainly strange, the HUMAN and AI_SHOOTER factors are never compared directly which results in unexpected behaviour we see here. Even if both terms were converted in the same fashion (5*factor (mod 65,535)), you might still see unexpected results.
i.e. HUMAN = 65,534 and AI_SHOOTER = 65,536 would result in AI not splitting even if it's stronger. 3DO error was to not compare directly HUMAN*CF to AI_SHOOTER which is what we would logically expect.

So how can one take advantage of this?
Simple, put more/fewer creatures in your largest stack. You can also try to play with your A/D primary skills as that can change the threshold for the splits.

Examples
I will do a few of the examples from the screenshot I posted at the start of the post.

Since only step 2 varies, I will do steps 1, 3 & 5 now.

1-
A/D (human) = 9/8 ... Strength_Human = 1.42478
A/D (AI) = 13/17 ... Strength_AI = 1.74714
CF = 1.42478/1.74714 = 0.81549

3-
32 Cyclops is 32*1266 = 40,512 (AI_SHOOTER)

5-
m = 1 + FLOOR(log(AI_SHOOTER,2)-16) = 1 + FLOOR(log(40,512) / log(2) - 16) = -1 -> always greater or equal to 1 so m = 1
(2^m = 2)

#####

Example A) 355 Champions

2- 355 Champions is 355*2100 = 745,500 (HUMAN)
4- HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*745,500*0.81549 (mod 65,535) = 25,093 (HUMAN_c)
6- 2^m*HUMAN_c = 50,186 > 40,512 = AI_SHOOTER
Therefore there is no split, AI deems human stronger.

Example B) 354 Champions

2- 354 Champions is 354*2100 = 743,400 (HUMAN)
4- HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*743,400*0.81549 (mod 65,535) = 16,531 (HUMAN_c)
6- 2^m*HUMAN_c = 33,062 < 40,512 = AI_SHOOTER
The human is weaker, let's split!

How many stacks?
FLOOR(AI_SHOOTER/HUMAN_c) = FLOOR(40,512/16,531) = 2
This falls in the category for 2 stacks.

Example C) 353 Champions

2- 353 Champions is 353*2100 = 741,300 (HUMAN)
4- HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*741,300*0.81549 (mod 65,535) = 7,968 (HUMAN_c)
6- 2^m*HUMAN_c = 15,936 < 40,512 = AI_SHOOTER
The human is weaker, let's split!

How many stacks?
FLOOR(AI_SHOOTER/HUMAN_c) = FLOOR(40,512/7,968) = 5
This falls in the category for 5 stacks.

Example D) 352 Champions

2- 352 Champions is 352*2100 = 739,200 (HUMAN)
4- HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*739,200*0.81549 (mod 65,535) = 69,941 (HUMAN_c)
6- 2^m*HUMAN_c = 129,882 > 40,512 = AI_SHOOTER
Therefore there is no split, AI deems human stronger.

Example E) 1 Champion

2- 1 Champion is 1*2100 = 2,100 (HUMAN)
4- HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*2,100*0.81549 (mod 65,535) = 8,562 (HUMAN_c)
6- 2^m*HUMAN_c = 17,124 < 40,512 = AI_SHOOTER
The human is weaker, let's split!

How many stacks?
FLOOR(AI_SHOOTER/HUMAN_c) = FLOOR(40,512/8,562) = 4
This falls in the category for 4 stacks.

#####

I have no doubt there is more to this type of splitting but it's interesting that this algorithm predicts the setup shown here.

It's a good start.

EDIT: resized picture, fixed () not showing properly, typos.

EDIT2: added a new step which permits comparison when AI_SHOOTER is larger than 2*65,535. It works for more cases but I seem to have some rounding issues or not exact figure at times.
The algorithm can be cleaned up some more as well.
____________
My Let's Plays: Metataxer's Revenge - The Empire of The World 2

 Send Instant Message | Send E-Mail | View Profile | PP | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted January 31, 2017 02:08 PM

Thank you for an interesting reading. I'll try to look into this problem myself too.

Fix your post please. Your pictures are too wide and stretch the forum.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
RoseKavalier
RoseKavalier


Admirable
Supreme Hero
posted February 02, 2017 03:21 AM
Edited by RoseKavalier at 00:13, 03 Feb 2017.

Thanks, I've resized the offending picture and fixed some other mistakes.

I've had a little bit more time to test the condition when AI_SHOOTER > 65,535. It behaves in a similar way (AI splits, doesn't split depending on HUMAN_c) which means that there is also some calculation being done on AI_SHOOTER - otherwise the AI would split every time once AI_SHOOTER > 65,535 since HUMAN_C cannot be greater than 65,535.

I'll fiddle with it a bit more to see if I can deduce the modification but your knowledgeable input would be most welcome!

########

Seems I've got the algorithm mostly working, was missing a magnitude order factor.

########

Let's see some more examples about the post that started all of this.



These 3 values will be constant throughout:
1-
Strength_AI = 5.95
3-
AI_SHOOTER = 9,000 * 517 = 4,653,000
Note the AI here is created with 3x9,000 Medusae stacks so that we consider 9,000 Medusae as its strongest shooter stack - not 27,000.
5-
m = 1 + FLOOR(log2(AI_SHOOTER)-16) >= 1
m = 1 + FLOOR(log(4,653,000)/log(2)-16) = 7 (m)

Example A


1-
Strength_HUMAN (A/D = 59/49) = ... = 3.691544...(Strength_HUMAN)

CF = 3.691544/5.95 = 0.620427629... (CF)

2-
HUMAN = 302 Arch Devils * 7,115 = 2,148,730 (HUMAN)

4-
HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*2,148,730*0.620427629 (mod 65,535) = 46,521 (HUMAN_c)

6-
is HUMAN_c * 2^m < AI_SHOOTER?
46,521*2^7 = 5,954,688 > 4,653,000
No - Human is stronger so we don't split. And so inverse occurs, AI combines its shooters because it fears your are stronger.

Example B


1-
Strength_HUMAN (A/D = 63/53) = ... = 3.891979...(Strength_HUMAN)

CF = 3.891979/5.95 = 0.654114106... (CF)

2- (same as Example A)
HUMAN = 302 Arch Devils * 7,115 = 2,148,730 (HUMAN)

4-
HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*2,148,730*0.654114106 (mod 65,535) = 15,221 (HUMAN_c)

6-
is HUMAN_c * 2^m < AI_SHOOTER?
15,221*2^7 = 1,948,288 < 4,653,000
Yes - Human is weak, let's split.

7-
How many stacks?
FLOOR(AI_SHOOTER/HUMAN_c) = FLOOR(4,653,000/15,221) = 306 <= 7
(# Stacks cannot be greater than 7.)
This falls in the category for 7 stacks.

Example C


1- (Same as Example A)
Strength_HUMAN (A/D = 59/49) = ... = 3.691544...(Strength_HUMAN)

CF = 3.691544/5.95 = 0.620427629... (CF)

2-
HUMAN = 17 Arch Devils * 7,115 = 120,955 (HUMAN)

4-
HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*120,955*0.620427629 (mod 65,535) = 47,539 (HUMAN_c)

6-
is HUMAN_c * 2^m < AI_SHOOTER?
47,539*2^7 = 6,084,992 < 4,653,000
No - Human is stronger so we don't split. And so inverse occurs, AI combines its shooters because it fears your are stronger.

Example D


1- (Same as Example A, C)
Strength_HUMAN (A/D = 59/49) = ... = 3.691544...(Strength_HUMAN)

CF = 3.691544/5.95 = 0.620427629... (CF)

2-
HUMAN = 16 Arch Devils * 7,115 = 113,840 (HUMAN)

4-
HUMAN_c = 5*HUMAN*CF (mod 65,535) = 5*113,840*0.620427629 (mod 65,535) = 25,467 (HUMAN_c)

6-
is HUMAN_c * 2^m < AI_SHOOTER?
25,467*2^7 = 3,259,776 < 4,653,000
Yes - Human is weak, let's split.

7-
How many stacks?
FLOOR(AI_SHOOTER/HUMAN_c) = FLOOR(4,653,000/25,467) = 91 <= 7
(# Stacks cannot be greater than 7.)
This falls in the category for 7 stacks.

########################

What can we conclude?

It works in very strange ways but also, it's possible to give a number of units to AI so that it's never willing to split. Or conversely, always willing to split because of the modulo operators.

There is also a similar phenomenon which happens with non-shooter stacks which can combine together if AI is stronger than you (speculating at this point, I have not investigated it yet). Maybe it will the feature of a future post.
____________
My Let's Plays: Metataxer's Revenge - The Empire of The World 2

 Send Instant Message | Send E-Mail | View Profile | PP | Quote Reply | Link
RoseKavalier
RoseKavalier


Admirable
Supreme Hero
posted June 03, 2017 03:59 AM bonus applied by angelito on 11 Jun 2017.
Edited by RoseKavalier at 16:37, 08 Jul 2017.

On diagonal movement discounts

Normally, diagonal step costs 141 movement (100×sqrt 2). However, it seens the developers thought it would be a nice feature to allow your very last step to be either horizontal or diagonal.

As stated on Tribute to Strategists p.12:
Quote:
“There is one peculiarity: if the hero has at least 100 movement points left he will still be able to make a diagonal move on non-penalty terrains.
This is not true for penalty terrains, even in case you have native creatures or (any level of) Pathfinding, you will not be able to move in any direction if you do not meet the movement requirement.”

What this should correctly say is:
[*]If you are granted non-penalty by creatures, you still need 141 movement for that last diagonal.
[*]The above only applies when you don’t have pathfinding. With pathfinding, you can take a diagonal step with 100 movement.

I eventually noticed that this was not always true, there are some occasions where the game lets you take a diagonal step without it costing 141 movement. I concluded that the movement algorithm was faulty and through some investigation managed to find it.



* Nomads grant temporary Pathfinding when on Desert and remove 50% penalty on Sand.



Some examples...
Example 1
Rough terrain with Gravel road. No pathfinding. Attempt 2 diagonals.
$movement = 181
$penalty = 125%
$feature = 65%

If we look in the Summary of bugs table, we see that 181 doesn't fall in the category for the bug. So we will get a full cost diagonal movement.
$cost=0.65×100×sqrt 2=91

The hero will be billed 91 movement which is expected with a diagonal step, leaving him with:
181-91=90 movement

Since the hero now has less than a diagonal movement on a Gravel road (91), he cannot take another diagonal. However he can travel one horizontal step (65) after which the game will round down movement to 0.



Example 2 (same as 1, except +1 movement)
Rough terrain with Gravel road. No pathfinding. Attempt 2 diagonals.
$movement = 182
$penalty = 125%
$feature = 65%

If we look in the Summary of bugs table, we see that 182 doesn't fall in the category for the bug. So we will get a full cost diagonal movement.
$cost=0.65×100×sqrt 2=91

The hero will be billed 91 movement which is expected with a diagonal step, leaving him with:
182-91=91 movement

This is enough for a second diagonal step which costs 91 movement points, like the first one.



Example 3 (same as 1 & 2, except less movement)

Rough terrain with Gravel road. No pathfinding. Attempt 2 diagonals.
$movement = 156
$penalty = 125%
$feature = 65%

If we look in the Summary of bugs table, we see that 156 falls in the category for the bug on rough terrain [125~176]. So we will get a cheap diagonal movement.
$cost=0.65×100=65

The hero will only be billed 65 movement which is surprising for a diagonal step, which leaves him with:
156-65=91 movement

91 movement is the minimum for a diagonal movement on gravel so the hero can take one more diagonal step.



Even though we had 26 fewer movement points in Example 3 than in Example 2 [which is what one would expect the minimum movement required for diagonals], we can still take 2 diagonal steps. Add insult to stupidity, in Example 1 we had 25 more movement points but the second diagonal step is impossible.

Enjoy this near useless information, except for the few puzzles out there

Some more examples
Cobblestone road (50% movement)
Non-native creatures
Swamp terrain (175%)
Basic pathfinding (25%)

>>Expected diagonal cost is 70, expected horizontal cost is 50.<<

25% pathfinding
Look up 175% - 25% range in table:
Min = 150, Max = 212

1) Movement = 247 (NOT IN RANGE)

2) Movement = 212 (IN RANGE)


DISCLAIMER: Thanks to AlexSpl for some verifications, see his post below for a few more examples.
____________
My Let's Plays: Metataxer's Revenge - The Empire of The World 2

 Send Instant Message | Send E-Mail | View Profile | PP | Quote Reply | Link
reald4rk
reald4rk

Tavern Dweller
posted June 04, 2017 12:26 AM

I want to make a tribute to Kicferk. Best homm 3 strategist.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted July 07, 2017 03:42 AM
Edited by AlexSpl at 18:38, 10 Jul 2017.

On diagonal movement discounts

2RoseKavalier: It seems H2 rules work for H3 as well. If a hero has enough movement points to make a horizontal step from the terrain* he's standing on, but doesn't have enough movement points to make a diagonal step from this terrain, he can make a step in any direction** at the cost of a horizontal step.

* Roads and Favorable Winds are not terrains.
** Note that Road and Favorable Winds' bonuses applied only when moving from one such tile to another.

* * *
Quote:
Conditions for bug to apply:
...
3) hero does not have pathfinding (Nomads included)
...

I suppose, it isn't correct. See the picture below. A knight (let it be Sylvia) has the standard army.



1. Sylvia has 210 MP and Basic Pathfinding

STEP 1
150 <= 210 <= 212 TRUE
The 1st step costs only 50 MP on Cobblestone road
a) 210 - 50 = 160

STEP 2
150 <= 160 <= 212 TRUE
The 2nd step costs 50 MP as well. Check it yourself. So, heroes with Pathfinding are also applicable, because 175 <= 160 <= 247 is FALSE, yet this step costs 50 MP
b) 160 - 50 = 110

STEP 3
150 <= 110 <= 212 FALSE
The 3rd step costs full 70 MP
c) 110 - 70 = 40

And Sylvia can't move even horizontally!

Well, you may say that 210 MP = 70 MP * 3 and my MP-meter lies? So, let us see what happens when...

2. Sylvia has 240 MP and Basic Pathfinding

STEP 1
150 <= 240 <= 212 FALSE
The 1st step costs 70 MP
a) 240 - 70 = 170

STEP 2
150 <= 170 <= 212 TRUE
The 2nd step costs 50 MP
b) 170 - 50 = 120

STEP 3
150 <= 120 <= 212 FALSE
The 3rd step costs 70 MP
c) 120 - 70 = 50

And now Sylvia can move horizontally (ofc, onto the road only)!

3. Sylvia has 240 MP and no Pathfinding

STEP 1
175 <= 240 <= 247 TRUE
The 1st step costs 50 MP
240 - 50 = 190

STEP 2
175 <= 190 <= 247 TRUE
The 2nd step costs 50 MP
190 - 50 = 140

STEP 3
175 <= 140 <= 247 FALSE
The 3rd step costs 70 MP
140 - 70 = 70

STEP 4!
175 <= 70 <= 247 FALSE
The 4th step costs 70 MP
70 - 70 = 0



Note that Sylvia with Basic Pathfinding can make only 3 diagonal steps, while without Pathfinding she can do whole 4. Weird.


* * *
In Russian:
Original puzzle
Solution

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
RoseKavalier
RoseKavalier


Admirable
Supreme Hero
posted July 07, 2017 02:40 PM
Edited by RoseKavalier at 14:47, 07 Jul 2017.

Hi Alex!

Thank you for checking this - it seems I forgot to check for Basic and Advanced Pathfinding and assumed they worked the same.

[...]
Edit: it seems you got it right, when you add levels of pathfinding, you need to look up values in the table which correspond to the new effective terrain penalty.
So if you're on 175% terrain with 25% reduction, look up 150% terrain.
[...]

I'll check it when I can and make the required modifications.
____________
My Let's Plays: Metataxer's Revenge - The Empire of The World 2

 Send Instant Message | Send E-Mail | View Profile | PP | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted July 08, 2017 05:23 PM

Quote:
**When $movement = 100 * sqrt(2) * $penalty, you technically have enough for a diagonal step when considering integer values - but you are still billed a horizontal step.

Indeed. And that's quite unexpected. So, the correct condition for "the edge effect" to appear is

Horizontal Step Cost <= Current Movement Pts <= Diagonal Step Cost

Quote:
when considering integer values

Movement points are always integers.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
RoseKavalier
RoseKavalier


Admirable
Supreme Hero
posted July 08, 2017 06:06 PM
Edited by RoseKavalier at 18:08, 08 Jul 2017.

Movement is indeed always integer, but I my inner thought is that ~technically~
100 * sqrt(2) = 141.42... > 141
Which may explain why developers made it as you wrote:

AlexSpl said:
Horizontal Step Cost <= Current Movement Pts <= Diagonal Step Cost

And that is what I have in the first picture, in mathematical terms.

Alex
____________
My Let's Plays: Metataxer's Revenge - The Empire of The World 2

 Send Instant Message | Send E-Mail | View Profile | PP | Quote Reply | Link
AlexSpl
AlexSpl


Responsible
Supreme Hero
posted July 08, 2017 06:14 PM

Quote:
100 * sqrt(2) = 141.42... > 141
Which may explain why developers made it as you wrote:

It makes sense

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
meKickfan
meKickfan

Tavern Dweller
posted August 06, 2017 11:12 AM

I WANT TO MAKE A TRIBUTE TO THE GUY THAT MADE KAPPA TOL RUSH..!! I WIN ALL MY GAMES NOW!!!!

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
spiryt
spiryt

Tavern Dweller
posted September 06, 2017 12:02 PM

Hello.

Sorry if already asked, but how is exactly random damage of creatures 'rolled' in games of Heroes series?

And how it can be replicated outside the game?

To me possibility of damage on all point of the scale seems about the same.

While calculating damage for each unit in the stack individually seems to produce results with a bell curve - way greater possibility of damage in middle zone...

I do suck at maths though.
____________

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
Orc
Orc


Famous Hero
posted December 21, 2019 07:45 AM

why does green dragons have better AI value than red dragons?

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
phoenix4ever
phoenix4ever


Legendary Hero
Heroes is love, Heroes is life
posted December 22, 2019 11:19 AM
Edited by phoenix4ever at 11:31, 22 Dec 2019.

That is honestly a real good question, since Red Dragons have 1 more attack, defense and speed than Green Dragons. They also cost 100 gold more.
I would suggest switching their AI Values.

 Send Instant Message | Send E-Mail | View Profile | Quote Reply | Link
Jump To: « Prev Thread . . . Next Thread » This thread is 20 pages long: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 · «PREV / NEXT»
Post New Poll    Post New Topic    Post New Reply

Page compiled in 0.1316 seconds