High Percentage misses

[hveldenz]

I cannot help but ask a few questions in regards to this thread, as it seems to be a recurring theme of prior threads (which is related to my continuing observation of the Uber ATGM sole survivor phenomenon). I have seen "99%" hit Hellfire B's in a solitaire campaign miss sometimes 4 times in a row - from a combination of evasion, arena fire, a non penetrating hit (actually a 'hit', but not the kill), or a straight out miss. I also seem to experience (with no testing like suggested) where the 66-69% numbers especially seem to come up the "Miss", or the "rally Failed" 8-10 times in a row.

I also do not believe there is intentional coding biases to the AI. However, if a developer can comment on :

1. Is the "to hit" number, or the rally percentage we see the actual final number rolled against by the random number generator? If not, it could be a partial explaination.
2. If the number %s are not what we the player sees - what goes on after the number displays, but before the die roll (Unit experience modifiers for example?)
3. How are random numbers generated in the game? The old days of coding random numbers in Basic or Fortran (from school days) involved selecting a seed number that directs into predetermined series of numbers. This essentially means that all random numbers are "predetermined", but not preselected until a seed number starts a chain of number results. The other method used in the past, IIRC, is a timing of a key press versus the computer clock speed to get to a seed or number. Does the random number algorithm have instead of true randomness like the perfect six sided die, just a large listing of numbers to read down in sequence (who remembers old PBM wargames using the stock tables and seed numbers to resolve a turn?)

I do appreciate the feedback from the programmers and developers, and also their collective effort.
Thanks.

[DRG]

When Andy gets back on line, which hopefully will be sometimes this week, I have him provide the exact details but ...

1] Yes a random number generator is used but he can provide more detail than I as to the order it is used

2] as above

3] as above.

I could give generalities but Andy can give the details so let's wait for the details. What I DO know and HAVE stated already on this forum to this same original poster is there is NO AI advantage built into the game. It uses the same code to determine it's hits and penetrations and it does your hits and penetrations. It's totally impartial in that regard and follows the same set of code rules. For anyone to suggest that we would code the game so that the AI is given the ability to shield it's "key" assets is insulting and absurd ESPECIALLY as I have already told anyone willing to listen that this does not happen. There is NO purpose in my lying about something like that. There is no benefit either personally or economically. The "profit" we make from this game wouldn't keep some people in beer for a year. Some people don't want to listen........ that's fine by me they can find something else to play.

Don

[Mobhack]

hveldenz said:
I cannot help but ask a few questions in regards to this thread, as it seems to be a recurring theme of prior threads (which is related to my continuing observation of the Uber ATGM sole survivor phenomenon). I have seen "99%" hit Hellfire B's in a solitaire campaign miss sometimes 4 times in a row - from a combination of evasion, arena fire, a non penetrating hit (actually a 'hit', but not the kill), or a straight out miss. I also seem to experience (with no testing like suggested) where the 66-69% numbers especially seem to come up the "Miss", or the "rally Failed" 8-10 times in a row.

I also do not believe there is intentional coding biases to the AI. However, if a developer can comment on :

1. Is the "to hit" number, or the rally percentage we see the actual final number rolled against by the random number generator? If not, it could be a partial explaination.
2. If the number %s are not what we the player sees - what goes on after the number displays, but before the die roll (Unit experience modifiers for example?)
3. How are random numbers generated in the game? The old days of coding random numbers in Basic or Fortran (from school days) involved selecting a seed number that directs into predetermined series of numbers. This essentially means that all random numbers are "predetermined", but not preselected until a seed number starts a chain of number results. The other method used in the past, IIRC, is a timing of a key press versus the computer clock speed to get to a seed or number. Does the random number algorithm have instead of true randomness like the perfect six sided die, just a large listing of numbers to read down in sequence (who remembers old PBM wargames using the stock tables and seed numbers to resolve a turn?)

I do appreciate the feedback from the programmers and developers, and also their collective effort.
Thanks.

1) it's the base to-hit before randoms f0r a few things that cannot be determined before the actual shot attempt (leader rating, random chance for a nearby HQ overriding the leader etc). (If it calculated the randoms, you would get a different to-hit every time you selected with the 'T' key, as you went back and forwards, and none would agree withthe rolled number. The actual number is posted on the shot attempt.

2) some random rolls for experience and chance - which are usually shown (depends on the delay) - missile supression etc seems to show more often for example.

3) was adressed earlier - and results from the RNG published, showing a flat distribution, and no "runs" of either high or low numbers.

4) the base to hit disregards stuff like ATGM avoidance, VIRSS and so on.

Cheers
Andy
(not on-line yet, just using a dial-up as some clown cut the Glenrothes blueyonder fibre optic cable the day BY installed my new kit - 49Kbips - we used to use that in 1998 .. !