Mully
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 18:26 |
|
One of the things we all get a chuckle at is how whene it seems like we've just rolled our 4th double skull of the game, we look up and see our "luck factor" is at 60%.
I'm curious to see if at the end of the match we can have a quick summary of our ALL our dice rolls for the game. Something real simple such as ;
1s 20%
2s 15%
3s 13%
4s 17%
5s 18%
6s 17%
Then we can see if all that bitching about skulls and missed dodges was legitimate or whining.
I know it would be extra work, but Christer looks so content sitting in front of his computer I figured he'd welcome the work. |
|
|
Mnemon
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 18:28 |
|
You'd have to ask SkiJunkie for that not Christer .
-Mnemon |
|
|
Christer
Joined: Aug 02, 2003
|
Well actually.. I created the luck calculation code.
-- Christer |
|
|
Mnemon
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 18:46 |
|
Yes - but I assume SkiJunkie would have to implement it in the client .
-Mnemon |
|
|
mamira
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 19:33 |
|
The problem is that these percentages don't translate well into luck.
Rolling a six when you only need a 2 is not really lucky. Rolling a six when you need it is ... |
_________________ I think, therefore I get paid. |
|
Mully
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 20:47 |
|
Quote: |
The problem is that these percentages don't translate well into luck.
Rolling a six when you only need a 2 is not really lucky. Rolling a six when you need it is ...
|
I understand that. I also understand that a failed pick up on turn 16 could be much more critical thana failed pick up on turn 2. Trying to interpret when the 1s come is impossible. I just though it would be interesting to see how my rolls averaged out in the game. OK - so it's not REALLY luck factor, "dice factor" instead. |
_________________ Owner of the REAL Larson
Come join the CCC League |
|
tgrozow
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 22:27 |
|
Since you roll quite a few dice every game, I expect to see a list like that look pretty much even in most of the games. I don't believe it will give you much information about how your die rolls were, unless they were exceptionally good or exceptionally bad.
I also don't think the luck factor is a good representation because in most games you look at the luck factor at the end and both teams would be somewhere around 50%, maybe a bit more or a bit less, but in a lot of games, luck is really one sided, as in one team makes everything, while nothing works for the other team.
I don't see any way other than going through the log file and analyzing to to figure out how lucky or unlucky a team was. |
|
|
Sinner
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 23:01 |
|
I think problem with this "luck-o-meter" is that it doesn´t reflect how early in a turn you fail something (just a guess). If this is right a failure with your first moved player counts as much as the failed action attempt with your last player.
So to correct this would be fairly easy. Just take the number of players that could take an action (even stunned do take actions, they roll over) per team turn, count the number of players that finished their move and if you fail a roll, you count the number of players left + the one that failed AS FAILED, because they didn´t get the chance to do anything so they failed their action
I know this is far from perfect, just a quick guess as I´m working on TSP algorithms atm, or should be at least, but I think it is more accurate than the actual calculation.
BUT: I don´t know if my assumption from the top is correct. |
_________________ Sinner
Darkie's Dreams - successfully cherrypicking any race, any coach, any rating, any number of DP since 20/09/2003 ... and still winning! |
|
Mnemon
Joined: Aug 02, 2003
|
  Posted:
Jul 16, 2003 - 23:56 |
|
No it isn't.
Say you prepared a two turn touchdown, with a goodnumber of players in position to score and your last move of the turn ends in your thrower failing the pickup. At that point your defense is most likely rather weak and you opponent might have a simple task at not only making your "almost certain" touchdown close to impossible but also will get in a good scoring position her/himself - i.e. a fail after moving most of your players can be just as bad as one on the first move.
-Mnemon |
|
|
Sinner
Joined: Aug 02, 2003
|
  Posted:
Jul 17, 2003 - 00:09 |
|
so it all gets down to what: the program can´t know which are "important" rolls and which are not.
But I still say that failing early in a turn is more dangerous to your drive than failing in the end of the turn. Just imagine: with Mnemon´s scenario from above: you have EVERY player except one exactly where you want him (well maybe not exactly but close enough), but if you fail your first action, you are doomed. Thats why blitzes and trapped wild animals can *beep* *bp* your tactics.
So no accurate luck calc, no whining about "bad luck" only thing that comforts me when I loose is that I AM a lousy coach, hehe.
I know and I believe.
G´Night |
_________________ Sinner
Darkie's Dreams - successfully cherrypicking any race, any coach, any rating, any number of DP since 20/09/2003 ... and still winning! |
|
Zhluhur
Joined: Aug 02, 2003
|
  Posted:
Jul 17, 2003 - 13:31 |
|
I want to know on how many GFI attempts I fail. So the Dice rolls have to be put in categorys.
Like:
GFI: 1 = 70%,
Skulls = 30% etc...
we want a total statistic of all actions or team does in a match! Per player!
I.E.
Player X has played 30 matches, rolled 23 Skulls, failed 15 GFI's and failed 10 dodges on a 1. |
_________________ *-* Let the games begin! *-* |
|
deneb
|
  Posted:
Jul 17, 2003 - 13:51Guest |
|
Hi All,
Have you got a document which explains how are calulated the "luck factor"?
If not? Can someone explain how is it working?
Thanks,
deneb |
|
|
deneb
|
  Posted:
Jul 22, 2003 - 13:10Guest |
|
Nobody knows?
even a quick one?
deneb |
|
|
Heruwulfar
Joined: Aug 02, 2003
|
  Posted:
Jul 22, 2003 - 13:54 |
|
deneb wrote: | Nobody knows?
even a quick one?
deneb |
me wondering too
|
|
|
Skar2
Joined: Aug 02, 2003
|
  Posted:
Jul 29, 2003 - 11:12 |
|
mamira wrote: | The problem is that these percentages don't translate well into luck.
Rolling a six when you only need a 2 is not really lucky. Rolling a six when you need it is ... |
If you ask a mathematician, this line doesn´t make any sense, because it´s not important what you need for a success by calculation probabilities. And therefore I don´t think this is so important.
What do you want as a luck counter? Is it the luck itself or a combination of luck and intelligence of the player? When you are interested in luck of a player (like the name of the counter) you should only take the statistic of the made rolls into account and then no value for successful or unsuccessful actions. An intelligent player would make the safest moves before the critical moves and therefore have more successful actions (and higher luck). But what the hell does that have to do with luck? So, if you ask me, the luck value should be calculated ONLY by statistical reasons.
But another thing would be interesting, too. How risky is a player. A player who takes great risks in his movements (like lots of GFI or dodges or else), will have a very low luck value, because most of his turns will end up in a failure. Perhaps there is a way to find an extra value for the risk a player used to play. There it came into account what rolls a player has to do. It could be calculated by the needed rolls he have to make (without taking the outcome into account). It isn´t important how often a 5+ roll was successful for the risk-value, it should be important how often one has taken the risk of a 5+ roll.
In my "dreams" of a perfect calculation there would be several interesting value, like the average players not moved due to turnovers, or number of missed GFI´s or so on. But I think it would be a great advantage to get two different values in the intention described above. Perhaps the risk-value could be coupled to each team or each coach, too, which could be interesting. And I would be very interested in the correlation between risk and success of a team/coach.
-------------------------------------
Some thoughts of a mathematician about luck in blood bowl |
|
|
|