pythrr
Joined: Mar 07, 2006
|
 Posted:
Jan 29, 2010 - 07:00 |
 
|
well, if the sun ain;t good enough, perhaps you could investigate Uranus...? |
_________________
 |
|
Christer
Joined: Aug 02, 2003
|
The blackbox system is tricky to figure out. So, let's see what the blackbox saw for that particular round:
12 people in total activated, and this is all the possible matchups the scheduler saw for each coach:
| Code: | Frankenstein
- Frankenstein vs Wakslo (Undead 138/149 vs 145/146 Chaos Dwarf) @989
- Frankenstein vs FMSTAN (Orc 126/106 vs 129/109 Khemri) @980
- Frankenstein vs Ropetus (Chaos Dwarf 171/162 vs 171/162 Skaven) @978
- Frankenstein vs Espionage (Orc 126/106 vs 124/108 Lizardmen) @976
- Frankenstein vs vicius (Dwarf 212/207 vs 205/197 Nurgle's Rotters) @975
- Frankenstein vs BillBrasky (Chaos Dwarf 171/162 vs 165/169 Chaos) @969
- Frankenstein vs Tueffy (Chaos Dwarf 171/162 vs 168/155 Dwarf) @937
- Frankenstein vs Slex (Chaos Dwarf 171/162 vs 205/177 Ogre) @932
- Frankenstein vs Shades_SteelFist (Orc 126/106 vs 102/97 Lizardmen) @914
- Frankenstein vs nimrodel (Orc 126/106 vs 100/92 Amazon) @902
- Frankenstein vs McCloud (Khemri 169/169 vs 201/183 Chaos Dwarf) @849
--
nimrodel
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
- Frankenstein vs nimrodel (Orc 126/106 vs 100/92 Amazon) @902
- nimrodel vs FMSTAN (Amazon 100/92 vs 129/109 Khemri) @900
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- nimrodel vs Tueffy (Amazon 100/92 vs 129/114 Skaven) @788
- nimrodel vs BillBrasky (Amazon 100/92 vs 116/110 Chaos Dwarf) @605
--
Wakslo
- Wakslo vs Slex (Chaos Dwarf 145/146 vs 144/149 Nurgle's Rotters) @992
- Frankenstein vs Wakslo (Undead 138/149 vs 145/146 Chaos Dwarf) @989
- Wakslo vs Ropetus (Chaos Dwarf 145/146 vs 137/134 Wood Elf) @928
- Wakslo vs BillBrasky (Chaos Dwarf 145/146 vs 156/162 Wood Elf) @927
- Wakslo vs Espionage (Chaos Dwarf 145/146 vs 142/150 Undead) @924
- Wakslo vs Tueffy (Chaos Dwarf 145/146 vs 168/155 Dwarf) @822
- Wakslo vs FMSTAN (Chaos Dwarf 145/146 vs 125/126 Necromantic) @694
--
McCloud
- McCloud vs Slex (Chaos Dwarf 201/183 vs 205/177 Ogre) @969
- McCloud vs Espionage (Chaos Dwarf 201/183 vs 189/182 Skaven) @921
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- McCloud vs Ropetus (Chaos Dwarf 201/183 vs 171/162 Skaven) @853
- McCloud vs BillBrasky (Chaos Dwarf 201/183 vs 156/162 Wood Elf) @852
- Frankenstein vs McCloud (Khemri 169/169 vs 201/183 Chaos Dwarf) @849
- McCloud vs FMSTAN (Chaos Dwarf 201/183 vs 141/169 Khemri) @834
--
Ropetus
- Ropetus vs Espionage (Chaos 136/133 vs 136/138 High Elf) @985
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- Frankenstein vs Ropetus (Chaos Dwarf 171/162 vs 171/162 Skaven) @978
- Ropetus vs FMSTAN (Wood Elf 137/134 vs 134/125 Norse) @978
- Ropetus vs Tueffy (Skaven 171/162 vs 168/155 Dwarf) @967
- Wakslo vs Ropetus (Chaos Dwarf 145/146 vs 137/134 Wood Elf) @928
- Ropetus vs Slex (Skaven 171/162 vs 205/177 Ogre) @923
- McCloud vs Ropetus (Chaos Dwarf 201/183 vs 171/162 Skaven) @853
--
Shades_SteelFist
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- Frankenstein vs Shades_SteelFist (Orc 126/106 vs 102/97 Lizardmen) @914
- Shades_SteelFist vs Espionage (Lizardmen 102/97 vs 124/108 Lizardmen) @894
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- Shades_SteelFist vs BillBrasky (Lizardmen 102/97 vs 116/110 Chaos Dwarf) @782
- Shades_SteelFist vs Tueffy (Lizardmen 102/97 vs 129/114 Skaven) @756
--
Tueffy
- Ropetus vs Tueffy (Skaven 171/162 vs 168/155 Dwarf) @967
- Tueffy vs BillBrasky (Dwarf 168/155 vs 165/169 Chaos) @946
- Tueffy vs Espionage (Skaven 129/114 vs 127/119 Wood Elf) @944
- Frankenstein vs Tueffy (Chaos Dwarf 171/162 vs 168/155 Dwarf) @937
- Tueffy vs FMSTAN (Skaven 129/114 vs 129/109 Khemri) @887
- Tueffy vs Slex (Dwarf 168/155 vs 205/177 Ogre) @887
- Wakslo vs Tueffy (Chaos Dwarf 145/146 vs 168/155 Dwarf) @822
- nimrodel vs Tueffy (Amazon 100/92 vs 129/114 Skaven) @788
- Shades_SteelFist vs Tueffy (Lizardmen 102/97 vs 129/114 Skaven) @756
--
vicius
- Frankenstein vs vicius (Dwarf 212/207 vs 205/197 Nurgle's Rotters) @975
- vicius vs Espionage (Nurgle's Rotters 205/197 vs 189/182 Skaven) @889
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- vicius vs FMSTAN (Nurgle's Rotters 205/197 vs 141/169 Khemri) @828
- vicius vs BillBrasky (Nurgle's Rotters 205/197 vs 218/216 Undead) @797
--
Espionage
- Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
- Ropetus vs Espionage (Chaos 136/133 vs 136/138 High Elf) @985
- Frankenstein vs Espionage (Orc 126/106 vs 124/108 Lizardmen) @976
- Espionage vs BillBrasky (Dwarf 119/113 vs 116/110 Chaos Dwarf) @968
- Tueffy vs Espionage (Skaven 129/114 vs 127/119 Wood Elf) @944
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
- Wakslo vs Espionage (Chaos Dwarf 145/146 vs 142/150 Undead) @924
- McCloud vs Espionage (Chaos Dwarf 201/183 vs 189/182 Skaven) @921
- Espionage vs Slex (High Elf 136/138 vs 144/149 Nurgle's Rotters) @920
- Shades_SteelFist vs Espionage (Lizardmen 102/97 vs 124/108 Lizardmen) @894
- vicius vs Espionage (Nurgle's Rotters 205/197 vs 189/182 Skaven) @889
--
BillBrasky
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- Frankenstein vs BillBrasky (Chaos Dwarf 171/162 vs 165/169 Chaos) @969
- Espionage vs BillBrasky (Dwarf 119/113 vs 116/110 Chaos Dwarf) @968
- Tueffy vs BillBrasky (Dwarf 168/155 vs 165/169 Chaos) @946
- BillBrasky vs FMSTAN (Wood Elf 156/162 vs 141/169 Khemri) @938
- Wakslo vs BillBrasky (Chaos Dwarf 145/146 vs 156/162 Wood Elf) @927
- BillBrasky vs Slex (Wood Elf 156/162 vs 205/177 Ogre) @898
- McCloud vs BillBrasky (Chaos Dwarf 201/183 vs 156/162 Wood Elf) @852
- vicius vs BillBrasky (Nurgle's Rotters 205/197 vs 218/216 Undead) @797
- Shades_SteelFist vs BillBrasky (Lizardmen 102/97 vs 116/110 Chaos Dwarf) @782
- nimrodel vs BillBrasky (Amazon 100/92 vs 116/110 Chaos Dwarf) @605
--
FMSTAN
- Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
- Frankenstein vs FMSTAN (Orc 126/106 vs 129/109 Khemri) @980
- Ropetus vs FMSTAN (Wood Elf 137/134 vs 134/125 Norse) @978
- BillBrasky vs FMSTAN (Wood Elf 156/162 vs 141/169 Khemri) @938
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- nimrodel vs FMSTAN (Amazon 100/92 vs 129/109 Khemri) @900
- Tueffy vs FMSTAN (Skaven 129/114 vs 129/109 Khemri) @887
- FMSTAN vs Slex (Khemri 141/169 vs 205/177 Ogre) @879
- McCloud vs FMSTAN (Chaos Dwarf 201/183 vs 141/169 Khemri) @834
- vicius vs FMSTAN (Nurgle's Rotters 205/197 vs 141/169 Khemri) @828
- Wakslo vs FMSTAN (Chaos Dwarf 145/146 vs 125/126 Necromantic) @694
--
Slex
- Wakslo vs Slex (Chaos Dwarf 145/146 vs 144/149 Nurgle's Rotters) @992
- McCloud vs Slex (Chaos Dwarf 201/183 vs 205/177 Ogre) @969
- Frankenstein vs Slex (Chaos Dwarf 171/162 vs 205/177 Ogre) @932
- Ropetus vs Slex (Skaven 171/162 vs 205/177 Ogre) @923
- Espionage vs Slex (High Elf 136/138 vs 144/149 Nurgle's Rotters) @920
- BillBrasky vs Slex (Wood Elf 156/162 vs 205/177 Ogre) @898
- Tueffy vs Slex (Dwarf 168/155 vs 205/177 Ogre) @887
- FMSTAN vs Slex (Khemri 141/169 vs 205/177 Ogre) @879 |
Ok, so now the scheduler looks for possible schedules by trying 10 different methods of pairing up people:
| Code: | Generating schedule #1
- Wakslo vs Slex (Chaos Dwarf 145/146 vs 144/149 Nurgle's Rotters) @992
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- Frankenstein vs FMSTAN (Orc 126/106 vs 129/109 Khemri) @980
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
- Shades_SteelFist vs Tueffy (Lizardmen 102/97 vs 129/114 Skaven) @756
New best score: 5511
Generating schedule #2
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- Tueffy vs Slex (Dwarf 168/155 vs 205/177 Ogre) @887
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- Wakslo vs FMSTAN (Chaos Dwarf 145/146 vs 125/126 Necromantic) @694
- vicius vs Espionage (Nurgle's Rotters 205/197 vs 189/182 Skaven) @889
- Frankenstein vs McCloud (Khemri 169/169 vs 201/183 Chaos Dwarf) @849
Generating schedule #3
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- Ropetus vs Tueffy (Skaven 171/162 vs 168/155 Dwarf) @967
- Espionage vs Slex (High Elf 136/138 vs 144/149 Nurgle's Rotters) @920
- vicius vs BillBrasky (Nurgle's Rotters 205/197 vs 218/216 Undead) @797
- Frankenstein vs McCloud (Khemri 169/169 vs 201/183 Chaos Dwarf) @849
Generating schedule #4
- Frankenstein vs vicius (Dwarf 212/207 vs 205/197 Nurgle's Rotters) @975
- McCloud vs Slex (Chaos Dwarf 201/183 vs 205/177 Ogre) @969
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- Ropetus vs Tueffy (Skaven 171/162 vs 168/155 Dwarf) @967
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
- Wakslo vs BillBrasky (Chaos Dwarf 145/146 vs 156/162 Wood Elf) @927
New best score: 5699
Generating schedule #5
- Wakslo vs Slex (Chaos Dwarf 145/146 vs 144/149 Nurgle's Rotters) @992
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- Tueffy vs BillBrasky (Dwarf 168/155 vs 165/169 Chaos) @946
- Frankenstein vs Ropetus (Chaos Dwarf 171/162 vs 171/162 Skaven) @978
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
Generating schedule #6
- McCloud vs Slex (Chaos Dwarf 201/183 vs 205/177 Ogre) @969
- Frankenstein vs vicius (Dwarf 212/207 vs 205/197 Nurgle's Rotters) @975
- Tueffy vs BillBrasky (Dwarf 168/155 vs 165/169 Chaos) @946
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- Wakslo vs Ropetus (Chaos Dwarf 145/146 vs 137/134 Wood Elf) @928
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
Generating schedule #7
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- Tueffy vs Slex (Dwarf 168/155 vs 205/177 Ogre) @887
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- Frankenstein vs Wakslo (Undead 138/149 vs 145/146 Chaos Dwarf) @989
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
Generating schedule #8
- Wakslo vs Slex (Chaos Dwarf 145/146 vs 144/149 Nurgle's Rotters) @992
- Frankenstein vs vicius (Dwarf 212/207 vs 205/197 Nurgle's Rotters) @975
- Tueffy vs BillBrasky (Dwarf 168/155 vs 165/169 Chaos) @946
- Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
- nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937
- McCloud vs Ropetus (Chaos Dwarf 201/183 vs 171/162 Skaven) @853
Generating schedule #9
- nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884
- Tueffy vs Slex (Dwarf 168/155 vs 205/177 Ogre) @887
- McCloud vs vicius (Chaos Dwarf 201/183 vs 205/197 Nurgle's Rotters) @865
- Frankenstein vs Wakslo (Undead 138/149 vs 145/146 Chaos Dwarf) @989
- Ropetus vs BillBrasky (Chaos Dwarf 153/157 vs 156/162 Wood Elf) @981
- Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
Generating schedule #10
- McCloud vs Slex (Chaos Dwarf 201/183 vs 205/177 Ogre) @969
- nimrodel vs FMSTAN (Amazon 100/92 vs 129/109 Khemri) @900
- Frankenstein vs Wakslo (Undead 138/149 vs 145/146 Chaos Dwarf) @989
- Shades_SteelFist vs BillBrasky (Lizardmen 102/97 vs 116/110 Chaos Dwarf) @782
- Ropetus vs Espionage (Chaos 136/133 vs 136/138 High Elf) @985 |
You'll see that schedule #4 is the one that was picked. Ok, so let's look at the two matches in particular:
| Quote: | [21:59] <BowlBot> - Shades_SteelFist vs FMSTAN (Lizardmen 102/97 vs 129/109 Khemri) @924
[21:59] <BowlBot> - nimrodel vs Espionage (Amazon 100/92 vs 124/108 Lizardmen) @937 |
The argument here is that looking at TS directly, it seems that the lizardmen teams should swap places here. Let's go back to the list of games and see what the suitability score was for those two potential games:
| Code: | Espionage vs FMSTAN (Lizardmen 124/108 vs 129/109 Khemri) @993
nimrodel vs Shades_SteelFist (Amazon 100/92 vs 102/97 Lizardmen) @884 |
A direct sum of the suitability scores show that the overall suitability score is indeed very slightly higher in the latter case (1861 in the chosen schedule vs 1877 in the swapped one).
The reason this schedule wasn't picked is that it simply wasn't ever considered. The scheduler has to be able to generate schedules quickly, and can't really do an exhaustive search of all possible schedules while still generating it in a reasonable time. In this case, it might have been possible (2 280 960 possible schedules), but the worst case scenario for 12 participants is closing in on 11! (which comes in at just under 40 million possibilities). With 20 participants in a round, it'll explode to a point where the servers will lock up for a considerable amount of time.
It could be argued that the chosen matchup would have been more fair than the suggested alternative, since the spread between the two suitabilities would have been higher. It's a matter of choosing between one match being fair and the other unfair and both matches being sort of in the middle ground.
You also have to understand that the scheduler is based on TS, which is focused on figuring out which of the teams is going to _win_ the game, rather than who does the most casualties. I'm pretty sure you wouldn't have been quite as vocal had it been a halfling team you were scheduled against for example.
In my opinion, the blackbox scheduler works quite well. It's not perfect though, because of the nature of the problem it tries to solve. But either way, if you feel that you are capable of coming up with a new system that is better than how the scheduler works I'll happily listen to your ideas. Now, the basic gripe you have isn't really the blackbox scheduler, but the TS system.
The TS system works better than any other metric on the site. The same offer stands for TS though: Come up with a better way to do what TS is designed for (the likelyhood of each team winning a match in TDs), feel free to get in touch. |
|
|
CircularLogic
Joined: Aug 22, 2003
|
| Posted:
Jan 29, 2010 - 12:23 |
|
I think one flaw that would be great to be tweaked is the evaluation of understrenght teams. A team with 8 players is overrated most of the time, which can lead to a bad vicious cycle.
Also one might think of weighting the impact of stars abit more. Most teams have a 'core team' of 4-8 players and the quality of that core is what is deciding. The rest is mostly bodies. That is also a reason, why hired stars are often underappreciated TS wise.
A more fundamental change could be switching the TS impact of the armor value from a sum to a product, because AV becomes more valuable, the better the player is. |
|
|
Christer
Joined: Aug 02, 2003
|
CircularLogic, please start a new thread about this and add some concrete examples on teams which would be more in line with real win probability with change like this. Try to apply the change to teams in various stages of development, and see what happens with the more extreme cases as well. Also, include your thoughts on how to identify the "core" group of players in a generic way. |
|
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jan 31, 2010 - 18:26 |
|
| Christer wrote: | | The TS system works better than any other metric on the site. The same offer stands for TS though: Come up with a better way to do what TS is designed for (the likelyhood of each team winning a match in TDs), feel free to get in touch. |
Basically, I'd suggest to just use LRB 5 TV instead.
That should account for all major flaws with the current TS-system.
You might want to keep the TS rules for FF/AC/CL/RR, regard handicaps as 5% rather than 5 straight TS and rework the TS for star players.
Definitely no discounts for injuries, exactly as with TV.
my2p |
|
|
shadow46x2
Joined: Nov 22, 2003
|
| Posted:
Jan 31, 2010 - 20:00 |
|
oh jesus christ...when will you people get it through your heads that LRB5 is *NOT* the fix to everything around here?
--j |
_________________
| origami wrote: | | There is no god but Nuffle, and Shadow is his prophet. |
  |
|
freak_in_a_frock
Joined: Aug 02, 2003
|
| Posted:
Jan 31, 2010 - 20:02 |
|
Personally i believe that many of the complaints about the formula for the box don't look at what it is trying to do.
The box is trying to make matchups that will produce a game that is the most likely to be fair, in terms of the chances of winning. Too often people disregard this and look at whether the formula produces games that will be fun or will result in one of the teams involved being bashed to a pulp.
The main proof i have for this is the continual belief that under-manned teams get drawn against ogre teams, and these matches aren't fair. Whilst i can agree that they are not desirable i can't agree that they are not fair. If this was the case ogres would have a far higher win rate than they do. As it is the box seems to over rate the ogres, producing games that they have little chance of winning. Quite often the games highlighted as unfair actually result in the underdog winning. |
|
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jan 31, 2010 - 20:20 |
|
| shadow46x2 wrote: | oh jesus christ...when will you people get it through your heads that LRB5 is *NOT* the fix to everything around here?
--j |
That's totally besides the point. The TS-formula has numerous horrible flaws and is illogical in many ways (I could provide examples, but really don't want to list details here).
TV doesn't have these flaws (except for DP, due to the different rules), is pretty much straightforward and easy to understand. |
|
|
Naru1981
Joined: Jan 01, 2009
|
| Posted:
Jan 31, 2010 - 20:40 |
|
I joined the box for the challenge. never knowing what your going to face. Where is the fun in a game you know you will proberbly win? We've all had games we dont think were fair but hey thats the fun of the box. We all love when we see we're facing flings and hate when we see Khemri etc. But again thats the fun of the box. |
_________________
 |
|
shadow46x2
Joined: Nov 22, 2003
|
| Posted:
Jan 31, 2010 - 20:59 |
|
way to cite opinion as fact without providing any examples...
nice troll, frankenstein...
--j |
_________________
| origami wrote: | | There is no god but Nuffle, and Shadow is his prophet. |
|
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jan 31, 2010 - 21:12 |
|
| shadow46x2 wrote: | way to cite opinion as fact without providing any examples...
nice troll, frankenstein... |
You can find 2 elaborated examples for horrible flaws in my blog.
Incidentally, if you intend to further insult me personally, I suggest that you continue that via PM, no need to do that in public, thx. |
|
|
shadow46x2
Joined: Nov 22, 2003
|
| Posted:
Jan 31, 2010 - 21:30 |
|
so you're citing a st2 troll slayer with MB as a reason why TS is borked?...
and TV doesn't have flaws of its own, right?
--j |
_________________
| origami wrote: | | There is no god but Nuffle, and Shadow is his prophet. |
|
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jan 31, 2010 - 21:46 |
|
| shadow46x2 wrote: | so you're citing a st2 troll slayer with MB as a reason why TS is borked?...
|
Firstly, that's a really horrible flaw there, indicating that something is fundamentally wrong with the TS-formula.
Secondly, you have entirely ignored the illogical TS for ogres.
Thirdly, I'm exclusively interested in a sober, factual analysis instead of cheap polemics, I guess we might have different priorities in that respect.
| Quote: | | and TV doesn't have flaws of its own, right? |
Namely? Now you have failed to provide any examples at all.
Anyway, this discussion is pointless, as your only interest is to goof me off regardless.
I was always under the impression that the TS-formula is more or less sacred, Christer's recent post suggested otherwise, so I was just trying to help out by providing a (not only in my opinion) better and more elegant way to go.
Shadow, I'm so not interested in any kind of dispute with a spiteful, unreasonable admin-watchdog (hadn't been back then with angie and are not, right now, with you), therefore I will cease and desist from posting in this thread any further, just spare the rest of the community of your animosity as well, ok? |
|
|
CircularLogic
Joined: Aug 22, 2003
|
| Posted:
Jan 31, 2010 - 22:58 |
|
SImple example why TV is much worse estimating winning prowess of a team:
A blodge elven lino has the same TV than one with passblock and diving catch. Now pitching two team against each other, one with 12 blodging proelf linos and one with 12 passblock diving catch proelf linos, each with 2RR. By my estimate, the blodging one will win more than 95% or the matches. |
|
|
pubstar
Joined: Jun 13, 2009
|
| Posted:
Jan 31, 2010 - 23:24 |
|
When evaluating a matchmaking system, I think 2 things are important to remember:
1. As Freak said, there's a difference between matches you can win, and matches you are happy to get.
2. It will never be perfect.
In 101 Box games (not a huge number, but enough for me to have formed an opinion), I've had maybe 5 matches that were truly unfair in terms of winning (probably fewer, but I may have forgotten a couple). There have been plenty of times when I didn't want a particular squad to face Khemri, but rare were the matches that I didn't have a fair chance to win (based on teams, anyway... dice and coach skill are another issue  ).
Maybe others have had vastly different experiences, but my personal Blackbox success rate is roughly 95%. So, accepting point 2 as a given... I don't see the system as a problem. Do you really want to wring hands and gnash teeth over that marginal 5% of bad matchups?
If there's some specific flaw you take issue with, it's certainly fair to bring it up. I thought Frankenstein's Ogre TS blog was well-presented, even if I don't agree. But to call for a complete removal and replacement of the current system seems inappropriate, as it seems to work at a fairly high rate. |
|
|
|
|
| |
Welcome Newcomers - ...
Search
