Fly Casual X-Wing Simulator

By s1ickrick, in X-Wing

19 minutes ago, tampermagnitude said:

I've seen some absolutely loony dice over a sustained period as well from this game, but I've seen the same sort of mad dice irl as well. It happens.

Like watching a focused A-Wing roll 5 blanks, twice in a row? Fortunately he was defending against single hits.

2 hours ago, Flurpy said:

Hi I am also a software engineer. Would you like me to independently check his code to verify the dice are random? Since the code is open source I can screenshot it for anyone interested here and explain how it works.

2 hours ago, Azrapse said:

You are blocking someone that has been exquisitely polite and trying to help?

Well, that solved the issue. You are the problem. You can't deal with the frustration of dice not going your way, and you are looking for excuses to explain why you lose or why the game doesn't go as you wish it went.

I have worked on Fly Casual's code. I have seen the dice randomization code. It randomly rotates the dice before rolling them, then they physically roll for extra randomness. There is no secret agenda, no bias towards the AI, nothing!

You aren't the first person I read complain about the dice being unfair. I have seen in in Vassal and in the other X-wing simulator that used to be around, "Squadron Benchmark".

It was always the same people that struggled dealing with frustration complaining about the dice benefiting the AI.

I've stated it before, but I'll say it again: I'm a software engineer with over 20 years of experience. I'm not involved in developing Fly Casual, but a while back, I did review the dice rolling code, including asking a few questions of Sandrem to get clarification on certain points. It works exactly as Azrapse is saying here. In fact, it randomly rotates the dice on three separate axes before it drops them and lets a physics engine roll the dice, so they're already random even before they roll. It's as random as any algorithmic random number generator can be. The same dice rolling code is used for both the player and the AI, so even if there was a bias, it would benefit the player just as much as the AI. Any perception that the behavior of the dice changes over time is entirely due to confirmation bias, and not due to the software.

Yes, the dice can go completely against you in a game. That can happen in real-life, too. It's just that the times the rolls don't go your way tend to stick in your memory more than the times that they do.

To be fair to everyone who brings this up periodically - I'm not immune that effect, either. I mean, there was a reason I was reviewing the code in the first place! 😅

Edited by Freeptop

Could always just go the way of, i think it was fire emblem that did this, and rolled the to hit ''dice'' twice, then took the result that had a better odd of happening. They did that as a way to placate all those persons who dont understands how odds work and who were furious over missing an attack with 80% chance to hit.

However, that means youll struggle to adapt to playing in real life, where the dice are actually random, and you prob wont improve as much.

3 hours ago, Wolfshead said:

Like watching a focused A-Wing roll 5 blanks, twice in a row? Fortunately he was defending against single hits.

Last week irl game, my jedi was behind an obstacle. An FO fired at it from range 3 and rolled two hits. I rolled 5 blanks. Blam, half points. Next round, that TIE FO fires range 3 at the same jedi. two hits, I roll 4 blanks. Dead jedi

Edited by wurms

Wait, some people are complaining that the AI is too hard to beat because of the random dice generator? Git gud applies a lot for that... My 10 year old is 50/50 so far, but he's very rusty at X-wing.

The only reason the AI has a shot is the built it squads working well for how the AI flies, if the dice were uneven (they're not), it'd add more challenge.

I'm not complaining at all, just recognizing what it can achieve at this point which is already quite impressive.

I've put it multiple games per day in the last few weeks, had a blast, became a patron and downloaded the source code to better understand why some glitches are happening. I'm looking forward to it getting even better.

But if you're losing too often, the dice are not the problem, you are. Git gud.

On 2/27/2020 at 1:24 PM, tampermagnitude said:

I've seen some absolutely loony dice over a sustained period as well from this game, but I've seen the same sort of mad dice irl as well. It happens.

100% agree but the dice on this are way more sustained than in real life. It has nothing to do with"git good". I mean you're going to take shots at some point (and don't really care to hear from the person that doesn't) and when the A.I. always seems to roll into full hits, skill has nothing to do with it. It just feels swayed in favor of the A.I.

24 minutes ago, Boba Fatt said:

...the A.I. always seems to roll into full hits, skill has nothing to do with it. It just feels swayed in favor of the A.I.

You hit the nail on the head with "seems" and "feels". It's a non-rational reaction. The code has been examined and explained. The problem lies with user perception. You might not like it, you might not believe it. But the problem is not in the game.

If it was just me saying this, I might believe you but others feel the same way (as noted in this thread). Also, the stats are right there on the page so my "feeling" and "seems" are justified.

Yes, other people feel the same way. Sometimes I feel that way too. We all have the same neurological hardware after all.

It just lies to us.

I play a lot of Blood Bowl online. The same tendency is prevalent amongst players. They just can't believe they should roll as many 1s in a sequence as they frequently do.

A physicist called Leonard Mlodinov has written a great book on the subject called "The Drunkard's Walk". In it he cites the famous example of the iPod's shuffle function. Users complained that it wasn't truly random because it kept repeating songs. But that's what random does. They had to rewrite it to make it *less* random. They dialed down the repetition to make it *feel* more random. To feel more random it had to be less random. Because humans are legendarily bad at judging probability. We have evolved to detect patterns everywhere, even when there is no pattern. It's why we see rabbits in clouds and Mother Teresa in a cinnamon bun .

I guess the bottom line is that if the perceived flaws in a games RNG is damaging your enjoyment you need to find one that is written better.

But I guarantee you that what you are really looking for is one that is written worse.

^^^

This. Congratulations sir, you win.

The probability code has been checked by at least 3 different sources, all saying the same thing (dice are unbiased, and even if they are, the players and the AI uses the same code to roll dice so at least both players should get the same advantage.) The creator of the program has been around and saying, often, that the code is unbiased, and really there would be no reason for him to ever ''cheat'' the code in favour of the computer. Not only that, but ''cheating'' the code in such a manner would be more work the just coding in at fair randomizer for the rolls, since it's already something that exists in C#.

And yet some people continue to think that there's a secret conspiracy intent of hiding the true nature of the program, which can distinguish when a player is playing the game and purposefully try to make him lose by having good rolls, but only, like... 20% more often that average.

And that point, what do we need to convince people that the whole thing is fair and it's just users perceptions that biased?

@Sandrem, I for one deeply enjoy Fly Casual, and I hope the conspiracy theorist won't dampen your enthusiasm! Continue on your awesome work please! :P

I'll never forget that keynote from Jobs where he talked about the iPod.

It really hit me because I was guilty of thinking the same thing. But the way he presented it hit the nail right on the head.

At the end of the day Fly Casual is a fantastic platform to get your fix if games aren't available or you can't escape from your family to play. At the very least you can get some practice in for positioning and getting used to a list. Help you learn the triggers etc. It will never replace a RL game. So even if you lose....so what? Fly casual man!

@Sandrem , Good job mate. Thank you for all your work and haters gonna hate. You've done what you can to explain. Leave it at that.

6 hours ago, Ghosthacked said:

At the very least you can get some practice in for positioning and getting used to a list. Help you learn the triggers etc. It will never replace a RL game.

Absolutely this. The dice are largely irrelevant to me tbh. They're just an entertaining sideshow that runs along while I'm gathering practical intel on my squad.

One thing about random is: THERE IS NO FAIR DIE. If dice where fair, we wouldn't need to roll them.

Another aspect of it is 'If my dice are bad, how well does this squad cope?

If it is 'Not well', now you are prepared... to maybe fly a bit better.

Or if it's 'Actually kind of ok.' Then maybe you're onto something.

Yeah honestly if the dice are really not in your favor then this honestly makes it an even better practice tool, you should never expect variance to bail you out in real life, this will just get you use to that

but in all honesty I really don’t notice these problems with the game, a lot of games I feel like I get amazing dice

On 2/27/2020 at 4:47 PM, Force Majeure said:

Additionally, a mode that happens to roll more criticals could be an interesting mode as well.

I believe there is already an option to make all hits roll as crits. because, you know, explosions :D

Oh dear gods, are we still talking about dice?

The other night at my kitchen table, my TIE/ba got killed at full health by a Range 3 unmodded shot that rolled hit, hit, crit into a Direct Hit. It was a 0.17% chance (or about 1 in 587).

Clearly, the dice and damage deck are coded to always play in my opponent's favor.

Clearly, better stop playing!

1 hour ago, StriderZessei said:

The other night at my kitchen table, my TIE/ba got killed at full health by a Range 3 unmodded shot that rolled hit, hit, crit into a Direct Hit. It was a 0.17% chance (or about 1 in 587).

Clearly, the dice and damage deck are coded to always play in my opponent's favor.

Just wait until it really happens in a tournament

The same phenomenon is known in computer games in general: if you play e.g. a RPG (roleplaying game) and the computer shows a probability of 80% that you hit something, players would be upset if it really was 80% - they would get the feeling that it is lower. So the game developers show 80% and internally they use 90% or whatever to please the gamers.

Why would Sandrem and the others try to cheat the players? What would be the motive? They want the people to play their game, and to enjoy it!

So, just to not talk about dice for a moment...

I'm relatively new to Fly Casual, and am thoroughly enjoying it (especially now that I know how to rotate my ships during setup. A whole new world!).

I'm particularly looking forward to flying a swarm of Fireballs, for no reason other than the phrase "swarm of Fireballs" is fun to say. I understand that implementation of pilots ships takes time, and am happy to wait for it to be done well, but was wondering how new pilots are implemented. Do new cards only get added with new releases, and I should be looking for 0.9.7. 2 , or are those changes made to an existing release, and a new download of the code will show the new pilots as they're added?

Again, I'm having a blast with the program. It's breathed new life into my household's interest in the game. Thanks for the hard work.

It is clear that the AI lobbyists have their hands in this. They are paying Sandrem big bucks under the table to make the dice favor the computer player. Corruption. Scandal.

But I'll probably keep playing Fly Casual. You know, because it is fun.

Edited by Tervlon

Couple small things :
When using the configuration with the vultures, if you are getting attacked through an asteroid, the attack should be obstructed and you should roll one extra defense die. But when you are attacking, you should ignore the rock (you are standing on, only, not any asteroid or debris).
However here you can see that the game only rolled 2 green dice for me when it should have been 3.
cLbgDUe.png

Something else i've noticed is that Avenger (Tie-Silencer pilot) ability always lets him do action even when stressed, such as after K-turning, when it should only be active when a friendly ship is destroyed.