kabill Posted July 24, 2013 Share Posted July 24, 2013 Also I think you have serious issues with you to hit probabilities. I don't know how you compute and apply them to the game, but after experimentation, I found that even if the to hit probability is represented as 70%, its more closely to 25% - just save a game, make a shot, record the outcome, and load back the game to repeat the shot, and so on... you might start looking at your subsequent shots not as independent event, but as a series of dependent events. How many shots have you tested doing this? You'd need a fair few to get something which is vaguely statistically accurate. My own experience of the RNG has been quite fair, so it might have just been a run of bad luck. Also, I don't think the game uses a seed for number generation (I've seen different results after loading before, without there being any changes made to what was done). Quote Link to comment Share on other sites More sharing options...
StellarRat Posted July 24, 2013 Share Posted July 24, 2013 (edited) Also, I don't think the game uses a seed for number generation (I've seen different results after loading before, without there being any changes made to what was done).Most of the software based random number generators use a seed. There is usually a function written by the compiler developers to generate randoms. So, you're really at the mercy of the compiler as to how good it is. The only question is whether or not the use the SAME seed when you load a save. Sounds like Xenonauts doesn't based on your comments.When we noticed problems with our game rolling extreme low probabilities too often we wrote a special program to actually studying the random number function. We ran thousands of tests. It turned out that it favored low numbers about 5 times as often as it should, but the rest of the numbers were pretty good. The solution was to flip the rolls so that low probabilities were determined by high rolls. That was a long time ago, but I'd be willing to bet not a lot of RD is spent on this particular subject. I think the computer lotteries actually use some type of hardware based physics to generate random numbers. Edited July 24, 2013 by StellarRat Quote Link to comment Share on other sites More sharing options...
kabill Posted July 25, 2013 Share Posted July 25, 2013 Most of the software based random number generators use a seed. There is usually a function written by the compiler developers to generate randoms. So, you're really at the mercy of the compiler as to how good it is. The only question is whether or not the use the SAME seed when you load a save. Sounds like Xenonauts doesn't based on your comments. Sorry, seed saving is exactly what I meant to say. Was having a brain-fail moment, apparently. The testing is interesting. Did the solution you mention actually get implemented? I have found myself surprised by the number of low-% shots hitting, but always figured that was just luck/bias on my part. Quote Link to comment Share on other sites More sharing options...
StellarRat Posted July 25, 2013 Share Posted July 25, 2013 Sorry, seed saving is exactly what I meant to say. Was having a brain-fail moment, apparently.The testing is interesting. Did the solution you mention actually get implemented? I have found myself surprised by the number of low-% shots hitting, but always figured that was just luck/bias on my part. I was talking about a game I developed many years ago with a couple of my buddies. We didn't sell many, LOL. But, that was one of issues we encountered in dealing with software generated random numbers. I have no idea if Xenonauts has any problems in that regard. I'm just saying that it's possible. Quote Link to comment Share on other sites More sharing options...
kabill Posted July 25, 2013 Share Posted July 25, 2013 Totally missed 'our' in your previous post. Oddly, the situation you described is how it felt to play the game today (more extreme high % shots missing than expected, more extreme low % shots hitting). But I was probably looking for it. Quote Link to comment Share on other sites More sharing options...
MrPyro Posted July 26, 2013 Share Posted July 26, 2013 I think the computer lotteries actually use some type of hardware based physics to generate random numbers. I've seen one of the original machines they used for the first computer-based lottery in the UK (Premium Bonds) in a museum. They state that it had a container of radioactive gas, and used the radioactive decay as part of their randomisation system, because radioactive decay is currently the only effect known that is completely random*. *Unless I'm behind the times, which is entirely possible. Quote Link to comment Share on other sites More sharing options...
StellarRat Posted July 26, 2013 Share Posted July 26, 2013 I've seen one of the original machines they used for the first computer-based lottery in the UK (Premium Bonds) in a museum. They state that it had a container of radioactive gas, and used the radioactive decay as part of their randomisation system, because radioactive decay is currently the only effect known that is completely random*.*Unless I'm behind the times, which is entirely possible. No, I think that is a way I heard of too. I'm sure it works well. Quote Link to comment Share on other sites More sharing options...
Mytheos Posted July 26, 2013 Share Posted July 26, 2013 Just as a side note, I tried one play through with using burst fire primarily on my rifles and another using mostly 20 TU shots. Maybe it was random but.... When using burst fire mostly, my Top kills were snipers with 15+ each followed by heavy with 7-8ish then the rifles were around 1-3. When using 20 TU shots mostly I had 1 riflemen constantly battling for the top kills spot with my sniper, and the ranges from top to bottom were MUCH more balanced, with my lowest kills around 7 and my top around 18 Using burst fire it was more 18-15-8-7-3-1-1-1, when using 20 TU shots it was 18-17-15-12-9-8-8-7. Quote Link to comment Share on other sites More sharing options...
svidangel Posted July 31, 2013 Share Posted July 31, 2013 Pointed this out in my drunken tactics thread a while ago, but 40TU burst fire has always been a terrible option unless your 20TU shot is over 95% accuracy. Basically unless you are jamming the weapon in their mouth, use single shots. The higher TU single shots will also do higher "average damage per round" at standard (non 95%) accuracy if you have to choose. Quote Link to comment Share on other sites More sharing options...
Lorebot Posted August 1, 2013 Share Posted August 1, 2013 I've never depended on burst fire for damage, even in EU1994. I use it mostly to cause suppression in Xenonauts since that can be more important than dealing damage given that most weapons simply aren't effective unless you can move in close. Being able to use 1 soldier to suppress a target, or group of targets, and then move another soldier in for the kill without fear of reaction shots is one of the main ways I keep my soldiers alive. Burst fire allows that from a much greater range than a grenade and is often much faster at suppressing a target than actually hitting them can be. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.