Insane Combat rolls - Can we get a replay feature

[KGB]

OK, I just got an insane combat loss.

I had my hero, a spider and a light cavalry and I attacked a L1 city with a catapult in it. So my stats were hero-6/spider-8/lt cavalry-4 against a catapult-5. I lost without doing a single hit!!!!!

The odds on that are approximately 150-1 given I didn't even hit the catapult one time. I think something was wrong with the combat numbers. I *suspect* the catapult was giving me a -2 to my stack because it was a city combat. But I can't be sure because the debug info doesn't show what the numbers were for the combat and there is no way to replay the battle.

Can we get a combat replay feature or have the debug logs at least show what the numbers were for neutral combats because some of the results I see are WAY outside what should be occurring or ever see occur with the actual Warlords 2 game. I should be able to play 100 more games this year and not see another lop sided result like that though I bet I do.

KGB

[piranha]

I will take some time the next week to make so all the dice rolls and strength are shown to the right when fighting so you know for sure if there was something wrong.

I will however still hold with my previous statment that the warlords system (at least warlords 1,2) does give you insane results now and then.
I don't like that part either.

I'd like to make some changes to the combat system. We might see if we can come up with some ideas for a improved combat system and present them and when we have something that should work we can build it into beta3.

[KGB]

Piranha,

I will take some time the next week to make so all the dice rolls and strength are shown to the right when fighting so you know for sure if there was something wrong.

Thanks. That will be very useful. It would be handy to have them in the debug logs too.

I will however still hold with my previous statment that the warlords system (at least warlords 1,2) does give you insane results now and then.
I don't like that part either.

Well, 150-1 works out to 99.3% chance for me to win the battle. How much more sure can one get? Results like that tend to take the strategy out of the game completely.

On the other hand, if the catapult was mistakenly giving me a -2 to my stack the odds for me to win would drop to 25-1 which is still 96% chance for me to win the battle.

I personally don't recall such crazy things happening in Warlords 2. I've played a dozen games with it in the last few months on Max Warlorders scenario he created and never saw anything like that. What happened to me was the equivalent of 8 light infantry attacking 8 light infantry and being left with 8 light infantry (instead of 1-2 light infantry).

I'd like to make some changes to the combat system. We might see if we can come up with some ideas for a improved combat system and present them and when we have something that should work we can build it into beta3

The thing is if you make many changes to the combat system then you no longer have Warlords. Not that that's bad thing, just that it's going to confuse a lot of player familiar with Warlords who are hoping to play a faithful remake.

My suggestion in the other thread of just throwing out the outlier results still makes the most sense. That lets you keep the same combat system but doesn't allow results like the one I experienced. Throwing out the outlier results would have had me lose the light cavalry and spider but my hero would have lived. It's still a combat that went against me but not one that was insanely against me.

KGB

[kenc80]

Sorry just saw this but wanted to comment because I agree: I'm seeing similar deviations.

As an avid and very loyal W2 player I do want to comment: my hotseat buddy and I are also seeing incredible rolls that we're simply not used to seeing in W2. Now, don't get me wrong, I had some big wins in W2 but alot of them we're due to ruin items and leveled up, blessed heroes.

I've had to be much more cautious in this game. Now is that bad? Well maybe not but its definitely a change. I talked about this in another KGB thread that I just saw a second ago --- if you change it --- thats fine --- but its a change --- and that may be good or bad, but I would love to see the combat system and game remain as close to W2 as possible --- but thats just my opinion (FWIW). I know that there are differences and you guys have done an awesome job --- I'm LOVING this. But I'm glad the conversation and the debate are going on.

KGB so glad you're involved and commenting, I remember debating wolves vs minos as best start units a coupla years ago with you on another Warlords board.

Take care

Ken

[KGB]

Ken,

I had another of the insane combat results today.

It's turn 13 in a FFA game where 2 players have already resigned and only 2 of us are left. Each with about 16-17 cities. My opponent with a slightly better position due to a rampaging hero stack that entered my land that I had no hope of stopping (oh, how I am DYING for the addition of the 'self raze my own cities' option in Beta 3 to deny his hero my cities...).

Anyway, his hero moving forward exposed his own cities which only had 1 light infantry in them. I have to make a move now or never. So I flung 3 Wizards down to rampage through his exposed cities. I reach the first city, attack and to my horror, all 3 wizards are killed by 1 light infantry. Let's see, he was strength 4 (+2 walls) but each Wizard is a 5. So his odds of winning over each wizard is < 40%. His chances of winning over 3 Wizards is <5%.

I resigned on the spot. A good game completely ruined by one outlier result.

These crazy results are reminding me of bad beats in poker where it has nothing to do with skill and all you can do is grit your teeth and hope you don't get bad dice in those key moments of the game.

KGB

[joasoze]

Hi - this is my first post here.

I would just like to add one thing. When you play as much as you do, you SHOULD see these results now and againg. This is excactly like the discussion when CIV IV was released.
If the odds of loosing a battle is 150/1 then you will experience such a loss if you play many games. If the odds of success is 99% then you will actually loos about 1 in a 100. Thats the way probability works and should not be removed from the game in my opinion. Its not insand but unlikely

[KGB]

Joeasoze,

Its not insand but unlikely

I realize you *should* see those results now and again. But those results are by nature unlikely (ie <5% is unlikely and <1% is extremely unlikely).

My point on that is:

1) It *should* never happen against neutral cities. As in the game should specifically prevent those kinds results against neutral cities. In player-vs-player it's a different story, but bad luck against the neutrals should not decide games.
2) *When* these insane results happen is far more important than the fact they can happen. If you have 8 men attacking a single unit and that unit wipes out 5 men, it's a crazy result but you still won so it's not a big deal. I also see crazy results at times when I lose battles I should have won but I have another stack ready to move in and attack so the crazy result isn't devastating. But when that result happens at a key moment in the game (hero vs hero battle, a battle like the one my Wizards lost where it would have won the game for me) it's not good because that one single battle turns an entire game.
3) It would be fairly easy to remove all the crazy results from the game entirely by chopping off the outer 10 or 20% of combat results so that only a result that fit in 80th or 90th percentile was used. That is by far the fairest thing that can be done. I say that because in a typical game there aren't enough combats to even out insane results. If every game featured at least 1000 battles then a couple of insane results aren't near as big a deal as a couple of insane results in a game with only 50 battles. In other words there isn't enough time for things to even out so the best thing is to remove them entirely.

Another thing I've noted recently as I pass the 60 games played mark. More and more it seems the insane results seem to be tied to specific days. As in I'm playing in 15 games at the moment. On days when I get a insane result in one game, I'll get it in the 3-4 other games I play a turn in that day. Otherwise, I don't see it in any games. I've learned now to play my turns first in games I am clearly winning in. If I see an insane result, I don't play turns in others specifically because it seems the dice rolls (ie random seed) are producing crazy stuff.

KGB

[joasoze]

I see your point. You want less randomness in multiplayer games. Still I disagree in your cure that would make neutrals pushovers and remove some of the charm in the game in my opinion. This is not chess. I guess we just have to agree to disagree.

[KGB]

Joasoze,

I don't want to make neutrals into pushovers. In fact one of the things I've asked for is the DLR feature of being able to have more than 1 unit in neutral cities.

What I want to remove is the case where 1 light infantry in a neutral beats 2 Minotaurs and Hero (Something I have experienced). If that 1 light infantry kills 1 or even 2 Minotaurs, that's more than enough randomness. It should not kill able to reach that <1% chance and kill all 3 units.

It could be another game option along with the size of dice, the combat system etc. Simply a "Smooth Combat Results" option that only allows results in the 90th percentile.

KGB

[Zajoman]

I agree with KGB completely.

How is an insanely extreme roll a charm in the game? When you experience three absolutely crazy losses in one day, I don't think you will feel charmed. Neither will new potential players. As KGB stated, results in 90th percentile are just about enough random. I vote for that.