vanGorn
Joined: Feb 24, 2004
|
 Posted:
May 25, 2007 - 22:20 |
 
|
I'm totally surprized about the thralls taking dodge as first choice on doubles.
Dozens of my thralls rolled doubles, so far, and I usually picked guard. That was based on the fact that strength skills are superior to the others and the official table: | Quote: | | (Vampire) Thrall: (44%)Guard , (22%)Leader , (19%)Pass , (6%)Mighty Blow, (3%)Dodge |
 |
_________________
Gimme a pint of fungus beer!
Then we will climb the ladder.
 |
|
Meech
Joined: Sep 15, 2005
|
| Posted:
May 25, 2007 - 22:46 |
|
vanGorn, looking at your only ranked Vampire team, you haven't rolled doubles with a thrall. Personally I agree with you that str skills are the way to go, but with mighty blow, because your opponent is less likely to be getting up when hit from a power punching thrall. |
_________________
Putting the FU in fumbbl since 9/2005 |
|
pac
Joined: Oct 03, 2005
|
| Posted:
May 25, 2007 - 22:54 |
|
| vanGorn wrote: | I'm totally surprized about the thralls taking dodge as first choice on doubles.
Dozens of my thralls rolled doubles, so far, and I usually picked guard. |
I was surprised too (and I just rolled a double on a Thrall and took Guard). But it's worked for him. 
If this project gets revised in the future, I would select the teams like this:
• Ranked teams (no change)
• At least 100 games (no change)
• Take the average win rate at the weighted TR for that team (using Race Stats and Rankings pages). Review the top three (or more) teams which are at least 5% but no more than 20% above that win rate.
So, for Humans the weighted TR is 181; in the 171-190 bracket, Human teams win 48% of games (rounding up). So Human teams which had won 53-68% of teams would be in.
But for now, I'm going to carry on and finish it off the way I have been doing it. |
_________________
Join us in building Blood Bowl Sixth Edition.
In other news, the Hittites are back. Join us in #fumbbl.hi  |
|
nin
Joined: May 27, 2005
|
| Posted:
May 26, 2007 - 20:51 |
|
| pac wrote: | | nin wrote: | | I have a cuestion, on the doubles information. Have you considered the (posible) difference on 1st skill roll being a Double and 1st Double rolled for a player. May be it's not relevant in those teams, I haven't checked it, but it's a relatively common cuestion in the forum |
Very good point. I've just finished Lizardmen - for whom Saurus are a case in point. I've had this in mind, but I haven't noticed it making a difference during the survey.
Of course, if a coach did turn down a double for Block (as with turning down stat ups), it's not easy for the survey to spot it. You could make a statistical estimate on it, but that still wouldn't tell you the conditions under which the stat was turned down. I think that sort of thing needs to be covered by another sort of guide. |
...and statisticks can lie humm, may be some of the reviewed teams coaches would like to make a comment
...and now that you write about Lizardmen (and Saurus) there's another cuestion on doubles:
rooky Saurus take the same Doubles as experienced ones?
I haven't played any long living Lizards, but I've had some succes with them (there and in TT) and for me it's that rooky or low experience Saurus take Frenzy, while more experienced ones take SF (if they have Breack Tackle) or Dodge (usually if they have both BT and Block)
may be it works for other players
may be it's excesive research for little info
...again, great work |
|
|
pac
Joined: Oct 03, 2005
|
| Posted:
May 27, 2007 - 14:50 |
|
High Elves are up: aka, 'What would kfoged do?' |
|
|
PurpleChest
Joined: Oct 25, 2003
|
Pac: I like the idea but i DO question your sample size.
To be clear there are two sample sizes here. You have ensured the player sample size ius large by taking long lived teams, thats fine.
But the play style sample size remains low as you have only chosen three teams. If these three coaches, or less, play in a similar stly, yet there IS a contrary stlye, will it be represented in the data? (eg: 3 elfballers, no battle elves) |
_________________
Barbarus hic ego sum, quia non intelligor illis -Ovid
I am a barbarian here because i am not understood by anyone |
|
Optihut
Joined: Dec 16, 2004
|
| Posted:
May 27, 2007 - 15:39 |
|
Sample size doesn't really matter - I won't let other people's choices influence my skill picks anyway, but for those that do struggle and can't make a decision, then looking at the most successful teams is exactly what they should do.
You have to look at the aim here, the goal isn't to cuddle everyone's ego and point out that sometimes pass block on a BOB can be a nice skill, but to provide a reference for newbies, who are interested in seeing what successful teams pick as skills. |
Last edited by Optihut on %b %28, %2007 - %18:%May; edited 1 time in total |
|
pac
Joined: Oct 03, 2005
|
| Posted:
May 27, 2007 - 16:14 |
|
| PurpleChest wrote: | | But the play style sample size remains low as you have only chosen three teams. If these three coaches, or less, play in a similar stly, yet there IS a contrary stlye, will it be represented in the data? (eg: 3 elfballers, no battle elves) |
In some cases, there has been a clear contrast in styles among the three teams (eg, Wood Elves). There, the 'minority opinion' is generally in there as a bracketed option.
I take the point you're making - something else to be considered in any second revision of the whole thing. (But for now I'll just try to complete it as it is now.) |
_________________
Join us in building Blood Bowl Sixth Edition.
In other news, the Hittites are back. Join us in #fumbbl.hi |
|
freak_in_a_frock
Joined: Aug 02, 2003
|
| Posted:
May 27, 2007 - 16:37 |
|
To be honest i am flattered to find that my necro team is the only one included in the guide. Shame no other long term necro teams seem to be able to see past a horde of dp zombies. Still we have a few more successful necro teams coming up behind me. |
|
|
pac
Joined: Oct 03, 2005
|
| Posted:
May 28, 2007 - 18:26 |
|
Just Ogres and Goblins to go now!
I'll probably do those two tomorrow. |
|
|
pac
Joined: Oct 03, 2005
|
| Posted:
May 28, 2007 - 20:16 |
|
Or I could just finish it off tonight!
I don't intend to do any more on this thing now. I'll leave that to some notional future reviewer - who I will point towards the points made earlier in the thread about things like adjusting the win% window for each roster. I would also suggest some kind of quick and easy system for checking range of opponents played.
There's every chance LRB 5 will have taken over before this wants doing again though … |
_________________
Join us in building Blood Bowl Sixth Edition.
In other news, the Hittites are back. Join us in #fumbbl.hi |
|
Optihut
Joined: Dec 16, 2004
|
| Posted:
Sep 22, 2007 - 18:38 |
|
I've just read through a couple of these - awesome work, well done pac!
I can see people going through the rosters in order to see what the popular choices are, but it's a great service to provide the results of the analysis for everyone. |
|
|
pac
Joined: Oct 03, 2005
|
| Posted:
Dec 12, 2007 - 14:56 |
|
I'm going to start work on revising this shortly, following the revised criteria below:
| pac wrote: | • Ranked teams (no change)
• At least 100 games (no change)
• Take the average win rate at the weighted TR for that team (using Race Stats and Rankings pages). Review the top three (or more) teams which are at least 5% but no more than 20% above that win rate.
So, for Humans the weighted TR is 181; in the 171-190 bracket, Human teams win 48% of games (rounding up). So Human teams which had won 53-68% of teams would be in. |
For a lot of teams this will make no difference. However, for teams with low win% (Vamps/Stunties/Ogres) it will lead to more teams being sampled.
If anyone would like to suggest any other criteria that should apply (maybe some kind of check that the team has played even-TS matches?), please do so quickly before I start!
For now, I'm going to work out the bracket for each list and include them below:
Roster - Avg weighted TR - Avg win% at that TR - Bracket
Amazon - 172 - 54.67 - 60-75%
Chaos - 193 - 49.99 - 55-70%
Chaos Dwarf - 181 - 51.49 - 56-71%
Dark Elf - 197 - 54.49 - 59-74%
Dwarf - 177 - 49.96 - 55-70%
Elf - 181 - 53.98 - 59-74%
Goblin - 159 - 48.17 - 53-68%
Halfling - 157 - 47.34 - 52-67%
High Elf - 187 - 52.06 - 57-72%
Human - 179 - 48.87 - 54-69%
Khemri - 174 - 46.92 - 52-67%
Lizardman - 173 - 50.77 - 56-71%
Necromantic - 168 - 48.31 - 53-68%
Norse - 165 - 46.11 - 51-66%
Nurgle - 175 - 47.72 - 53-68%
Ogre - 182 - 43.62 - 49-64%
Orc - 193 - 51.20 - 56-71%
Skaven - 178 - 53.69 - 59-74%
Undead - 177 - 51.70 - 57-72%
Vampire - 170 - 47.56 - 53-68%
Wood Elf - 189 - 56.78 - 62-77% |
|
|
PhoenixxX
Joined: Jan 29, 2007
|
| Posted:
Feb 12, 2008 - 14:38 |
|
I have a suggestion to the even matchup case:
Multipy the number of games vs a race by the STR diff given in the 'show stats' (it is already averaged).
Add all the numbers and divide them by the played games.
Now you have an average STR diff for all their games.
I would suggest a range of +/-5 but at higher TR it could be difficult to find even oppos, so maybe +/-10.
I made this for 2 of the Amazon teams:
Aggressive Asian Avengers: 10.9
One Night Stand: 0.9
A famous HE team for comparison Nandorins: 7.9
Also as a criteria to not reward cherrypickers:
Max: 40-60% elfes&skaven&stunties played
This should be lower(40%) for basherraces and higher(60%) for the 3 mentioned categories.?
Hard to find stunties with 100+ games who plays dorfs and orcs only 
Maybe there also should be a minimum to avoid teams that never choose tackle, because they never play elves?
But that hardly is believable ;o)
Same 3 teams again:
Aggressive Asian Avengers: 42.8%
One Night Stand: 46.8%
Nandorins: 55.9%
Note: I for myself, prefer to play av7 and elven teams, so i know how hard it is to play bashers a lot with this teams.
But, as I understand, this is a guide for newbies. Cherrypickers options may not the correct picks for them as newbies mostly play all races with their teams (eg. Dorfs taking MB always as 1st skill and then only playing elves). |
|
|
pac
Joined: Oct 03, 2005
|
| Posted:
Feb 12, 2008 - 14:50 |
|
Yep, those look like good solutions.
To be honest, when considering the possibility of a review of this, I've come up against the fact that these teams we may or may not disapprove of really don't pick skills any differently.
As with the argument made by pmg earlier in this thread that I should use teams coached by the coaches he rated most highly for reference, a look at the teams in question just doesn't reveal any notable difference in skill preferences from the ones already listed.
So, the net result of a review, in my opinion, would be a lot of work for someone, a new set of team names in the headers, and pretty much the same skill progression below.
All the best to anyone who wants to do it, but I decided against! |
Last edited by pac on %b %12, %2008 - %15:%Feb; edited 1 time in total |
|
|
|
| |
Gnomes are trash
Search
