Azure
Joined: Jan 30, 2007
|
 Posted:
Nov 09, 2012 - 17:33 |
 
|
I would be interested just in deviation from norm - the easiest and generally (yes - generally not all the time) means that armor/injury rolls are key.
I would be interested in:
Armor rolls made:
Expected # of armor breaks:
Actual # of armor breaks:
And then:
Injury rolls made:
Expected # of KO:
Expected # of CAS:
Actual # of KO:
Actual # of CAS:
I think this would serve 2 purposes - (1) a very easy way to check to see if someone did indeed get lucky with their armor/injury rolls...and (2) to help understand whether a bash first strategy is working for a team...if you are on average for armor/injury - yet are losing...maybe you will see a need to change strategies. Helps people rely on strategy instead of just luck to win. |
|
|
Overhamsteren
Joined: May 27, 2006
|
| Posted:
Nov 09, 2012 - 17:38 |
|
| Barre wrote: | | Overhamsteren wrote: | | Maybe something like average number of players on the field for the combined game so if you lose a lot of players early it will be a lower number than if you lose a lot of players in the late game. |
This can be determined already, sum the turns for all players then divide by 16 (or game turns for OT)
|
True a lot of things can already be calculated, but statistics should be easy and quick to read. |
_________________
Like a Tiger Defying the Laws of Gravity
Thanks to the BBRC for all the great work you did. |
|
freak_in_a_frock
Joined: Aug 02, 2003
|
| Posted:
Nov 09, 2012 - 17:42 |
|
Don't do it!!! (unless it is the gobbos eaten thing) |
|
|
PainState
Joined: Apr 04, 2007
|
| Posted:
Nov 09, 2012 - 17:43 |
|
My issue with stats as related to BB is that I agree that one "bad" roll could turn an entire game and the significance of that failed roll does not show up in the stat sheet as a glaring blunder.
BUT
Lets revist that Tunrover discussion. IF you tracked a team over 30 games just looking at that stat. And the team was 8/4/18 over 30 games and averages over 7 Turnovers a game. Well that stat is a big indicator of why a team is probally losing.
During that same run he had 6 TO in seperate games, on double skull. Lost 4 of those matches. Well you could say those were key moments but then you have to find out when they happened, what was the situation. DId the other team exploit it and get a TO and so forth. So the actual stat line of # of Double skull as an example needs further examination to determine the impact on the actual match.
I used the Stats for the Buccs to look at trends over a long span of matches. To attempt to help me play by looking at things Iam doing bad or good. And as a mental defense of how bad NUFFLE has a grudge against me. |
_________________
Comish of the:  |
|
happygrue
Joined: Oct 15, 2010
|
| Posted:
Nov 09, 2012 - 17:56 |
|
I agree with Azure, that would be great. I'll add that I would like to see more expected value computations, especially on dodges/leaps but other things would be great too. Quick summary for anyone who doesn't know what I'm talking about, if you did an action with three 3+ dodges (without the dodge skill) your expected number of dodges would be 3*(4/6)= 2. So at the end of a match if you got a summary like this:
EV dodges made: 25.2
Actual dodges made: 46
you could get a sense of how lucky or unlucky you had been on dodging. That would be great for passes, leaps, handoffs, wild animal, etc. But if I could pick one thing it would be dodges. Of course it would get a bit more complicated checking for tackle, using RR and such, but even a simple check would be great.
One other thing would be expected/actual RR use. I'm thinking team RR here, but skill RR would also be interesting. So if you used 4 team RR it would be nice to see how much that SHOULD have helped you (rerolling double skulls and such) and how much it actually did. I think that would be more useful even than counting turnovers, because so many turnovers are noncrital things like end of turn dodges and stuff that you kind of expect to fail anyway.
So a summary like:
EV Team RR success (5 used): 1.7
Actual team RR success: 4
would show (to some extent) that you were RR lots of high risk stuff and it paid off for you. Though one tricky spot that comes to mind would be RR block dice, as what is a success there depends quite a bit on skills. Sometimes you just need a push (when surfing or with strip ball) and sometimes you want a pow.
EDIT: Also, and this may be pie in the sky, but I would be really interested in looking at things like the breakdown of EV of success per action of a coach and also for a given team. That is, to get a ballpark of the risk that a given team is taking. So those Skaven with the great winning record after 100 games, what is that coach doing to get there? Is he dodging and making lots GFI, or -2d on the ball? Or running his players around with move actions a lot but rolling fewer dice than I am? When I watch the great coaches play they often take risks that I might not take, but when I look closely I see how high the payoff can be and just wonder about playstyles and if they are taking more risks or just BETTER risks. Anyway, not a high priority thing, but it would be great if you could wave a magic wand and make it so.  |
|
|
Ehlers
Joined: Jun 26, 2006
|
| Posted:
Dec 20, 2012 - 17:54 |
|
So was searching for something and found something that the Cyanide client had and just wanted to ask if we can also get that?
Ball possession %
Occupation of opponent half %
Occupation of own half % |
|
|
TheSyron
Joined: Jun 17, 2012
|
| Posted:
Dec 20, 2012 - 18:29 |
|
Something akin to the bbmanager floating around for the Cyanide game would be great. Pretty much everything it does has already been suggested, I think. It's nice to see after a match that it just wasn't your imagination that you had incredible/terrible luck (in general). |
|
|
JackassRampant
Joined: Feb 26, 2011
|
| Posted:
Dec 20, 2012 - 18:30 |
|
These stats I think are all relevant:
Actions lost to turnover
Actions declined
Go For It attempts
1/3d or 1/2d blocks, 1d blocks, 2d blocks, 3d blocks*
2+ skilled rolls, 2+ unskilled, 3+ skilled, 3+ unskilled, etc.*
* Or a matrix of rolls by odds, indicating number made
Knockdowns, knockdown %
Expected knockdown %
Drive chart: turn of kickoff, kicking team, drive outcome, number of turns, value on pitch at start of each drive
Re-Rolls used
Re-Roll success rate (how often RR generated categorically better outcome than original roll)
Expected Re-Roll success rate
% Re-Rolls used on non-turnovers
Pass attempts
Ejections
Cas caused by fouls
Surfs
Cas caused by surfs
KOs
KO recovery luck
Value on pitch, on average
In particular, success rate by difficulty, TRR success rate, lost actions, and drive charting would be really helpful. |
_________________
Lude enixe, obliviscatur timor. |
|
stej
Joined: Jan 05, 2009
|
| Posted:
Dec 20, 2012 - 19:25 |
|
As I've said in several threads before, I'd love to see something that shows you, broken down by each roll typ, the target value of the roll, the number of attempts at that target value and the number of successes or for something like block dice the result chosen.
Would give you an idea if you are making too many risky plays or if you have just got snaked a lot! |
|
|
Chainsaw
Joined: Aug 31, 2005
|
| Posted:
Dec 20, 2012 - 19:50 |
|
Really, what we all want and need and really deserve is beat 'em up style combo bling with a dramatic voiceover.
"QUAD SKULLS! BWUHAHAHAHAA!"
"11 2 plus combo!"
"12 2 plus combo!"
"13 YOU DIRTY ELF LUCKER!" |
_________________
Coach Chainsaw's Dugout
Free Gamer - blog - community |
|
Panda_
Joined: Jul 14, 2004
|
| Posted:
Dec 20, 2012 - 20:55 |
|
Turnovers can be measured by the number of players that didn't move yet.
Sum of those numbers each time a turnover happens.
- It leads new players to use as much players as possible before the turn overs (which will increase their play)
- It is an hidden indicator of someone's "luck"
|
_________________
"Rien ne sert de partir a point, il vaut mieux courir." |
|
fidius
Joined: Jun 17, 2011
|
| Posted:
Dec 20, 2012 - 21:11 |
|
While I can appreciate why you would not want to revisit the real-time Luck-O-Meter, Kalimar, I'd love to see a visual presentation of the ups and downs of a match in graphical form after the fact. Like Azure, my main interest is in deviation from norms. To this end I started developing (and never finished) a "luck index" to track successes and failures.
The simple idea was that over an infinite sample the Luck Index should equal 0. So if a 2+ roll fails 1/6 times, if the failure is worth -1.00, the success should be worth +0.20.
Also I wanted to build in the relative risk of having rerolls (or reroll skills) in-hand. So a success with RR in your pocket is worth less than one without. But also the net index gain or loss after roll plus re-roll should net to the same change as the theoretical "2+ w/ RR" chance. One would think that because a 1/36 fail is 1/6 as likely as a 1/6 fail, it should hit you harder, say -6.00. Unfortunately the math doesn't work that way; you can't expand both sides of the equation. So I took the SQRT: sqrt(6)=2.45. I know this is mathematically imperfect, open to input on how to improve this.
Anyway the simple base case looks like this:
| Code: | Roll Undesired Perms Success Fail
2+ no RR 1 6 +0.20 -1.00
2+ #1 w/ RR 1 6 +0.07 -0.35
2+ #2 w/ RR 1 6 +0.42 -2.10
2+ w/ RR 1 36 +0.07 -2.45
|
Then you just expand it for 3+, 4+, 5+ and 6+ rolls.
(To illustrate the imperfection noted above, here is the 6+ table. Ideally the 6+ no-RR fail would give you an equivalent Index change as a 2+ success, but as you can see, it doesn't, quite.)
| Code: | Undesired Perms Success Fail
6+ no RR 5 6 +2.24 -0.45
6+ #1 w/ RR 5 6 +1.66 -0.33
6+ #2 w/ RR 5 6 +2.00 -0.40
6+ w/ RR 25 36 +1.66 -0.73 |
The model applies to any roll that is d6-related, including armour rolls:
| Code: | Undesired Perms Success Fail
test vs AV5 10 36 +0.22 -0.56
test vs AV6 15 36 +0.36 -0.51
test vs AV7 21 36 +0.65 -0.47
test vs AV8 26 36 +1.15 -0.44
test vs AV9 30 36 +2.14 -0.43
test vs AV10 33 36 +4.59 -0.42
MB vs AV5 6 36 +0.13 -0.64
MB vs AV6 10 36 +0.22 -0.56
MB vs AV7 15 36 +0.36 -0.51
MB vs AV8 21 36 +0.65 -0.47
MB vs AV9 26 36 +1.15 -0.44
MB vs AV10 30 36 +2.14 -0.43 |
The Index rises and falls as successes and failures are made. Then you overlay a line chart of each coach's rolls over time and see visually how their relative luck impacted the game.
Obviously everything gets more complicated with block dice and injury because there are varying definitions of success/fail (defdown v push v bothdown v attdown; stun v KO v Cas). Also simply counting knockdowns is one thing; counting 1d naked Pows into blodge (my personal pet peeve) is another.
Not sure this is the type of feedback you were looking for K, but there it is. Likely it's ground you covered in the luck-o-meter originally... |
|
|
|
|
| |
War Drums?
Search
