Can you not post it anymore so that it at least looks fair Kennedy Parker !
-
LIVE: Richmond v Melbourne - 7:25PM Wed
Squiggle tips Demons at 77% chance -- What's your tip? -- Team line-ups »
Sim Sim update planning
- Thread starter Kennedy Parker
- Start date
- Tagged users None
- Jul 16, 2015
- 92,358
- 157,307
- AFL Club
- Richmond
- Other Teams
- Fighting Furies
- Moderator
- #152
What are you talking about?
- Jun 29, 2021
- 3,916
- 8,930
- AFL Club
- Richmond
I thought I might see you wander in here
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #154
Mobbs
Turns out coding in interchange substitutions was way easier than I thought. Looks like we won't need to kick Headless out of the league after all.
Just three layers if IF statements, 5% chance of interchange happening any given transaction, from there another roll to determine which bench player gets put on, one more IF to check which team is in posession and I found a dictionaries can swap keys within three lines.
Simmed a test game using Round 1 Royals and Furies lineups. After manually calculating the DT (haven't coded that in yet), we can see its relatively even compared to the current sim. The most consistent premium positions seem to be centre and full forward, but otherwise, any position can get any stat category, e.g. followers can take marks, a forward can lead the tackle count.
Turns out coding in interchange substitutions was way easier than I thought. Looks like we won't need to kick Headless out of the league after all.
Just three layers if IF statements, 5% chance of interchange happening any given transaction, from there another roll to determine which bench player gets put on, one more IF to check which team is in posession and I found a dictionaries can swap keys within three lines.
Python:
def Interchange():
SubstitutionEvent = random.randint(1,20)
if SubstitutionEvent == 1:
WhichBench = random.randint(1,2)
if sim_game.possession == "Home":
if WhichBench == 1:
KeyON = "h_INT1"
KeyOFF = random.sample(list(Player.homePos_Players), 1)
KeyOFF = KeyOFF[0]
print("ON: " + Player.homePos_Players[KeyON], KeyON)
print("OFF: " + Player.homePos_Players[KeyOFF], KeyOFF)
Player.homePos_Players[KeyON], Player.homePos_Players[KeyOFF] = Player.homePos_Players[KeyOFF], Player.homePos_Players[KeyON]
elif WhichBench == 2:
KeyON = "h_INT2"
KeyOFF = random.sample(list(Player.homePos_Players), 1)
KeyOFF = KeyOFF[0]
print("ON: " + Player.homePos_Players[KeyON], KeyON)
print("OFF: " + Player.homePos_Players[KeyOFF], KeyOFF)
Player.homePos_Players[KeyON], Player.homePos_Players[KeyOFF] = Player.homePos_Players[KeyOFF], Player.homePos_Players[KeyON]
elif sim_game.possession == "Away":
if WhichBench == 1:
KeyON = "a_INT1"
KeyOFF = random.sample(list(Player.awayPos_Players), 1)
KeyOFF = KeyOFF[0]
print("ON: " + Player.awayPos_Players[KeyON], KeyON)
print("OFF: " + Player.awayPos_Players[KeyOFF], KeyOFF)
Player.awayPos_Players[KeyON], Player.awayPos_Players[KeyOFF] = Player.awayPos_Players[KeyOFF], Player.awayPos_Players[KeyON]
elif WhichBench == 2:
KeyON = "a_INT2"
KeyOFF = random.sample(list(Player.awayPos_Players), 1)
KeyOFF = KeyOFF[0]
print("ON: " + Player.awayPos_Players[KeyON], KeyON)
print("OFF: " + Player.awayPos_Players[KeyOFF], KeyOFF)
Player.awayPos_Players[KeyON], Player.awayPos_Players[KeyOFF] = Player.awayPos_Players[KeyOFF], Player.awayPos_Players[KeyON]Simmed a test game using Round 1 Royals and Furies lineups. After manually calculating the DT (haven't coded that in yet), we can see its relatively even compared to the current sim. The most consistent premium positions seem to be centre and full forward, but otherwise, any position can get any stat category, e.g. followers can take marks, a forward can lead the tackle count.
| HO | K | M | HB | T | FF | FA | G | B | DT | |
| Engimal v3 | 0 | 4 | 1 | 1 | 2 | 0 | 0 | 0 | 0 | 25 |
| Blacky | 0 | 10 | 2 | 6 | 3 | 0 | 0 | 0 | 0 | 60 |
| PMBangers | 0 | 10 | 3 | 5 | 1 | 1 | 0 | 2 | 0 | 66 |
| Smartys Power | 0 | 3 | 1 | 4 | 4 | 2 | 0 | 0 | 0 | 38 |
| boncer34 | 0 | 4 | 2 | 8 | 3 | 1 | 2 | 0 | 0 | 41 |
| MrPremiership | 0 | 5 | 0 | 5 | 1 | 1 | 4 | 0 | 0 | 18 |
| krakouers85 | 0 | 5 | 5 | 1 | 4 | 0 | 1 | 0 | 0 | 45 |
| nahnah | 0 | 14 | 3 | 9 | 5 | 5 | 1 | 0 | 0 | 91 |
| Fitzey | 0 | 3 | 2 | 3 | 1 | 1 | 1 | 1 | 0 | 29 |
| ShaunDuggan | 0 | 8 | 7 | 3 | 2 | 2 | 0 | 2 | 1 | 74 |
| CakeEater | 0 | 1 | 4 | 7 | 2 | 2 | 0 | 0 | 0 | 39 |
| Ulahoopski | 0 | 1 | 1 | 4 | 1 | 0 | 0 | 0 | 0 | 18 |
| MC Bad Genius | 0 | 9 | 4 | 2 | 0 | 1 | 0 | 4 | 3 | 71 |
| The Filth Wizard | 0 | 10 | 4 | 1 | 0 | 0 | 0 | 1 | 1 | 51 |
| Dancing Potato | 0 | 7 | 2 | 1 | 3 | 4 | 5 | 2 | 1 | 43 |
| GreyCrow | 19 | 4 | 2 | 6 | 2 | 2 | 0 | 1 | 0 | 65 |
| spookism | 1 | 6 | 1 | 2 | 3 | 4 | 0 | 1 | 1 | 49 |
| Marlowe | 0 | 6 | 3 | 4 | 2 | 1 | 0 | 1 | 0 | 50 |
| grumbleguts | 0 | 5 | 2 | 7 | 2 | 1 | 0 | 1 | 0 | 50 |
| Paddles_ | 0 | 2 | 2 | 3 | 2 | 1 | 1 | 0 | 0 | 24 |
| rfctigerarmy | 0 | 2 | 0 | 4 | 1 | 1 | 4 | 0 | 0 | 7 |
| Ned_Flanders | 0 | 3 | 0 | 4 | 1 | 0 | 2 | 0 | 0 | 15 |
| Edmund Hunt | 0 | 4 | 1 | 4 | 1 | 0 | 1 | 0 | 0 | 24 |
| damicky | 0 | 3 | 0 | 2 | 3 | 1 | 0 | 0 | 0 | 26 |
| tigland | 0 | 5 | 5 | 6 | 0 | 1 | 2 | 1 | 0 | 43 |
| Tiger2709 | 0 | 9 | 3 | 4 | 2 | 0 | 5 | 0 | 0 | 37 |
| crazywildhorse | 0 | 1 | 0 | 5 | 3 | 0 | 0 | 0 | 0 | 25 |
| Elton Johns Wig | 0 | 10 | 4 | 12 | 1 | 2 | 0 | 1 | 3 | 81 |
| Sterge | 0 | 5 | 0 | 9 | 0 | 3 | 1 | 1 | 0 | 39 |
| Brunswick Trap King | 0 | 4 | 4 | 4 | 2 | 4 | 2 | 1 | 0 | 44 |
| GREENESHOOTS | 6 | 6 | 2 | 3 | 3 | 1 | 1 | 2 | 1 | 59 |
| Rioli8217 | 0 | 3 | 3 | 10 | 7 | 0 | 7 | 0 | 1 | 46 |
| Lynchy04 | 0 | 4 | 3 | 4 | 5 | 0 | 0 | 2 | 2 | 63 |
| Bastyy | 0 | 11 | 7 | 2 | 2 | 0 | 0 | 5 | 3 | 99 |
| TJASTA | 0 | 6 | 5 | 5 | 2 | 0 | 0 | 1 | 1 | 58 |
| U2tigers | 12 | 7 | 1 | 3 | 0 | 0 | 0 | 2 | 0 | 54 |
| Tigerturbulance | 0 | 8 | 3 | 6 | 1 | 1 | 0 | 1 | 1 | 57 |
| Mister M | 0 | 6 | 3 | 2 | 1 | 0 | 0 | 0 | 1 | 36 |
| Jackbero | 0 | 7 | 1 | 10 | 5 | 1 | 3 | 0 | 1 | 57 |
| Wacky Tiger | 0 | 6 | 3 | 6 | 6 | 0 | 1 | 0 | 0 | 60 |
- Moderator
- #155
This is really good.
- Jun 18, 2016
- 51,637
- 98,985
- AFL Club
- West Coast
- Other Teams
- Perth Scorchers
In breaking news, U2tigers has announced his retirement
Ophidian Old Boys
On second thought, just how good a Qooty player could I be
Metalcrusher
A God Amongst The People
- Banned
- #159
Mobbs
A Large Portion, Yes
Looks great! I really hope this can lead to a full changeover. A sim of this kind for aussie rules made from current day coding tools would be a real leap forward!
- Jun 18, 2016
- 51,637
- 98,985
- AFL Club
- West Coast
- Other Teams
- Perth Scorchers
Kennedy Parker, do you think there's scope to be able to produce stats by position (ie. a player changes positions and stats can be extracted for both positions) and head-to-head stats?
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #162
To have a complete report at the end of the game for each player's stats separated into each position they played. It would be possible but I think its a lot of work for very marginal benefit. The stats data overwrites each time it updated so we'd need to record a new set of stats concurrently each time there's a rotation.
Something more manageable would be to print a player's stat line at the time of the position change into a csv. Then with some subtraction in excel you could calculate their stats by position. e.g. if a player spent half the game at wing and were benched at half time, their stat line could be printed into csv, then when they come back on at full back until the end of the game, you could subtract end of game stats with stats at time of change to derive a breakdown of their game by position if you wanted to.
Head-to-head stats are probably possible, but I don't have any plans on implementing that. Don't know how the current Mobbs Medal votes are calculated but I plan on making it a head-to-head measurement.
- Jun 18, 2016
- 51,637
- 98,985
- AFL Club
- West Coast
- Other Teams
- Perth Scorchers
A .csv at the time of changing position could be a manageable middle ground
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #164
Dropbox - File Deleted - Simplify your life
First crack at producing the commentary file. It's extremely buggy at the moment as I haven't spent too much time on it, I think a lot of the issues are caused by boundary throw-ins which is annoying as I thought I had it working in an earlier version, but its not registering anymore after I added in individual players.
Anyway the main thing is for the time being, the scoreboard appears to be updating correctly.
Broken
Brownlow Medallist
Awesome stuff KP.
The voting is probably the most obvious thing needing a revamp. I don't know if this helps, or hinders, but the idea I was toying with was adding a new player attribute for BOG calculations, each event in the game sim would trigger a new calculation for BOG points (i.e. a goal to put the team in the lead late in the game might be 10 points, a junk time goal might be 1 point,...) and the players with the most BOG points would get the votes.
There was also the hope to make it position dependent if I got it working well (which may have fed in to Barrybrans idea of breaking down stats vs position).
But like I said, if this helps, bonus, if not, I'm sure you'll nail it.
- Jun 18, 2016
- 51,637
- 98,985
- AFL Club
- West Coast
- Other Teams
- Perth Scorchers
Dropbox - File Deleted - Simplify your life
www.dropbox.com
First crack at producing the commentary file. It's extremely buggy at the moment as I haven't spent too much time on it, I think a lot of the issues are caused by boundary throw-ins which is annoying as I thought I had it working in an earlier version, but its not registering anymore after I added in individual players.
Anyway the main thing is for the time being, the scoreboard appears to be updating correctly.
The great part is having an updated sim gives us scope to improve features along the way. I think this is a brilliant start.
I notice the quarter times are different too KP. Is there a randomiser to determine quarter length?
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #167
To my knowledge, the current sim bases Mobbs Votes off an attribute called 'form' that does not get posted to the forum but is in the sim files. I'd make a similar measure, but also factor in the performance of direct opponents. i.e. if a player gets +3 points for taking a mark, the oppo might get -2 points for conceding a mark etc.
Weighting points depending on the context of the game sounds cool, but I think it would require too many conditionals to adapt to all the possible contexts. The only way you can define the criteria would be with margin at a stage of the game, e.g. Team A comes back from 8 goals down at half time, a player who kicks the first 2 goals to kickstart it would get few points as they could be considered 'junk time' but a player who kicks the last one to put the team in front would get 5x the points even though all goals were equally important in the team winning.
Yeah, quarter end time can be anywhere between 25minutes and 35minutes, probability of quarter ending progressively increases until it is 100% after 35 minutes.
Riviat
Brownlow Medallist
Some of the bog scoring is starting to creep into supercoach scoring territory. + points for an effective kick vs ineffective
Mobbs
A Large Portion, Yes
Just on this (somehow I wasn't getting alerts for this thread), I was trying to create a tool which could do this, but I didn't complete it.
I can extract all the data for a chosen player separated by their opponents including time spent versing each other, but I got distracted and didn't add functionality to also extract the data for the opponents of that chosen player (so it's just one side of the comparison).
I'll have another look at it someday but it got hella confusing.
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #170
After a painstaking troubleshooting session with plenty of trial and error. I finally managed to run a test sim that actually properly included boundary throw ins that reflect on the stat sheet. Didn't know what exactly was wrong before and don't know why what I did fixed the issue but in the end it did.
The sim produced a total of 68 hitouts as opposed to the usual no. of goals + 4 for each quarter we are used to.
| HO | K | M | HB | T | FF | FA | G | B | DT | D | |
| Engimal v3 | 0 | 5 | 0 | 2 | 4 | 0 | 2 | 0 | 0 | 29 | 7 |
| Blacky | 0 | 6 | 2 | 6 | 3 | 0 | 2 | 0 | 0 | 42 | 12 |
| PMBangers | 0 | 2 | 1 | 10 | 2 | 1 | 1 | 0 | 0 | 35 | 12 |
| Smartys Power | 0 | 3 | 0 | 3 | 1 | 2 | 1 | 0 | 0 | 18 | 6 |
| boncer34 | 0 | 4 | 2 | 5 | 3 | 0 | 0 | 0 | 0 | 40 | 9 |
| MrPremiership | 0 | 3 | 1 | 4 | 0 | 2 | 6 | 0 | 0 | 4 | 7 |
| krakouers85 | 0 | 5 | 1 | 2 | 0 | 2 | 1 | 0 | 0 | 21 | 7 |
| nahnah | 0 | 12 | 3 | 9 | 5 | 7 | 3 | 0 | 0 | 81 | 21 |
| Fitzey | 0 | 4 | 3 | 1 | 4 | 1 | 1 | 0 | 0 | 37 | 5 |
| ShaunDuggan | 0 | 4 | 0 | 3 | 2 | 3 | 2 | 0 | 1 | 24 | 7 |
| CakeEater | 0 | 8 | 6 | 3 | 2 | 3 | 1 | 3 | 3 | 77 | 11 |
| Ulahoopski | 0 | 4 | 1 | 3 | 0 | 2 | 0 | 1 | 0 | 29 | 7 |
| MC Bad Genius | 0 | 6 | 4 | 5 | 1 | 0 | 2 | 1 | 2 | 46 | 11 |
| The Filth Wizard | 0 | 7 | 6 | 3 | 1 | 1 | 0 | 0 | 3 | 53 | 10 |
| Dancing Potato | 0 | 7 | 4 | 2 | 2 | 1 | 2 | 2 | 2 | 54 | 9 |
| GreyCrow | 7 | 3 | 1 | 2 | 3 | 1 | 0 | 0 | 0 | 36 | 5 |
| spookism | 0 | 8 | 5 | 6 | 1 | 2 | 0 | 0 | 0 | 57 | 14 |
| Marlowe | 0 | 3 | 1 | 8 | 2 | 2 | 0 | 1 | 0 | 44 | 11 |
| grumbleguts | 0 | 2 | 0 | 2 | 0 | 2 | 0 | 0 | 0 | 12 | 4 |
| Paddles_ | 34 | 5 | 0 | 3 | 0 | 2 | 0 | 0 | 0 | 57 | 8 |
| rfctigerarmy | 0 | 1 | 1 | 2 | 1 | 1 | 0 | 0 | 0 | 15 | 3 |
| Ned_Flanders | 0 | 7 | 2 | 6 | 5 | 2 | 2 | 0 | 1 | 56 | 13 |
| Edmund Hunt | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 0 | 7 | 2 |
| damicky | 0 | 2 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 14 | 3 |
| tigland | 0 | 3 | 3 | 8 | 2 | 0 | 6 | 0 | 0 | 24 | 11 |
| Tiger2709 | 0 | 5 | 0 | 5 | 2 | 1 | 1 | 0 | 0 | 31 | 10 |
| crazywildhorse | 0 | 7 | 2 | 9 | 2 | 0 | 1 | 0 | 0 | 50 | 16 |
| Elton Johns Wig | 0 | 13 | 6 | 10 | 4 | 4 | 5 | 2 | 1 | 95 | 23 |
| Sterge | 0 | 3 | 4 | 3 | 1 | 1 | 3 | 1 | 0 | 29 | 6 |
| Brunswick Trap King | 0 | 7 | 3 | 8 | 4 | 4 | 3 | 0 | 1 | 58 | 15 |
| GREENESHOOTS | 0 | 5 | 1 | 10 | 4 | 0 | 1 | 2 | 0 | 63 | 15 |
| Rioli8217 | 0 | 5 | 0 | 4 | 2 | 1 | 3 | 0 | 0 | 23 | 9 |
| Lynchy04 | 9 | 7 | 3 | 2 | 1 | 2 | 2 | 1 | 1 | 50 | 9 |
| Bastyy | 0 | 7 | 4 | 4 | 2 | 2 | 1 | 2 | 0 | 60 | 11 |
| TJASTA | 0 | 5 | 1 | 7 | 1 | 1 | 1 | 0 | 1 | 35 | 12 |
| U2tigers | 3 | 9 | 4 | 1 | 2 | 1 | 2 | 2 | 1 | 60 | 10 |
| Tigerturbulance | 0 | 2 | 2 | 12 | 3 | 2 | 0 | 0 | 0 | 50 | 14 |
| Mister M | 0 | 2 | 1 | 2 | 0 | 1 | 2 | 0 | 0 | 8 | 4 |
| Jackbero | 0 | 5 | 3 | 1 | 1 | 0 | 0 | 2 | 3 | 45 | 6 |
| Wacky Tiger | 15 | 3 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 30 | 3 |
The sim produced a total of 68 hitouts as opposed to the usual no. of goals + 4 for each quarter we are used to.
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #171
From a sample of 3 games, there definitely seems to be a high level of variance in this sim. All three games were simmed with no weather feature, meaning the pace of the game was exactly the same and produced drastically different results.
Game 1
====================
END OF QUARTER 4
GCR: 6.21.57
FUR: 6.13.49
====================
Game 2
====================
END OF QUARTER 4
GCR: 13.14.92
FUR: 10.18.78
====================
Game 3
====================
END OF QUARTER 4
GCR: 13.15.93
FUR: 16.16.112
====================
Even within games the pattern of the game can swing drastically, for instance, in Game 3, the Furies were only on 9.12.66 at three quarter time, but kick 7.4 in the final quarter. This happens in the current sim as well, but usually would require weather conditions conducive to high scoring.
Interestingly enough, average DT across games 1 and 2 were both 44. Average disposals both around 10.4.
And game 3, despite the drastically higher scoring had the same number of disposals and only 2pts higher average DT.
Game 1
====================
END OF QUARTER 4
GCR: 6.21.57
FUR: 6.13.49
====================
Game 2
====================
END OF QUARTER 4
GCR: 13.14.92
FUR: 10.18.78
====================
Game 3
====================
END OF QUARTER 4
GCR: 13.15.93
FUR: 16.16.112
====================
Even within games the pattern of the game can swing drastically, for instance, in Game 3, the Furies were only on 9.12.66 at three quarter time, but kick 7.4 in the final quarter. This happens in the current sim as well, but usually would require weather conditions conducive to high scoring.
Interestingly enough, average DT across games 1 and 2 were both 44. Average disposals both around 10.4.
And game 3, despite the drastically higher scoring had the same number of disposals and only 2pts higher average DT.
Mobbs
A Large Portion, Yes
I jiggled around stuff for years trying to get scores to both be sane but also variable, and same with players stats. Every tweak tends to effect everything else and the perfect balance is really tricky (if achievable), no matter what. I think you will be faced with this same challenge regardless of what approach or tools you use, and it just takes trial and error and testing a bunch of scenarios out, being ready to backtrack every time.
Riviat
Brownlow Medallist
That's some serious innaccuracy. If that happened in real games, the FF would be dumped every other week.
Kennedy Parker
🎨𝖆𝖗𝖙𝖎𝖘𝖙🎨
- Thread starter
- Moderator
- #174
If I were to introduce weather, I'd be fairly content for it to only have a marginal effect tbh and not define the pattern of the entire game
Riviat
Brownlow Medallist
Game needs an all-in brawl chance at half-time which triggers a mini game that is played out like what happens in PhenomenalV1's wrestling sims. Something different and maybe a bit of a laugh if it triggers but odds would be very low.
