New match report page (with some dice stats)

[Gandalf]

The continuing downtime of BB Manager (3-4 weeks now?) has spurred me on to create some kind of dice breakdown for us all to use. So here it is, the new re-designed and upgraded match reportpage.

The stats are very factual (ie you rolled a 2 for dodge), so doesn't rely on the complicated not-yet-fully-tested logic that you can see in the match log, which will eventually give you chances of success, re-roll statistics etc. etc... the aim was very much something quick to produce, and accurate.

There may be some dice roll type it is missing entirely (eg I was missing bonehead rolls as the log called them bone-head), if you notice any then let me know. Also I have just noticed Abom rolled 1 on a 2D6.... working theory is that all 2D6's are 1 too low, I'll get to this at some point. I think that's the limit of what can go wrong?!

Credits: The general layout of the page is something I discussed with our "man-on-the-street" Daigaro a few months back, I think it broad terms it works well (maybe less so in the detail). The block dice analysis is inspired by something la told me. General comments on the layout are welcome. "Put lines in between the rows & columns" is probably the next thing to do - it's not super-readable at the moment. Specific things around creating new information that isn't available elsewhere will be met with a "go give me feedback on the match log first".

One final limitation. The match log needs to be viewed at least once (by anyone, at any time) before the dice stats will appear. This limitation will go once the log files & save files parts of the program are fully integrated. And hey, there's a link to view the match report on the match detail page now, so it's a lot easier

[notjarvis]

Looks really good. Good stuff.


Quick question - what does Defender Stumbles(+BD) mean?

[Gandalf]

(+BD) means that both down was rolled as well. Sometimes this is a good result - even better than Defender Stumbles, depending on the skills involved. So I wanted to provide some visibility of that. For example a new Norse team against a new Amazon team would mean Defender Stumbles(+BD) is always going to be better than just Defender Stumbles.

[Raveen]

Awesome! This to be honest was the last missing link in BBDB for me. Great work on getting it done, even if it's a bit crude!

Could you add lines at the bottom of the table showing totals of each dice result (so 20 1's, 13 2's etc.) and below that a line showing the proportion of the total (30% 1's, 15% 2's etc.) please? It's easily done in Excel but would be good to have on the page itself. Also table lines but you're onto that already.

BBM produces a graph of dice results, is there any desire for that?

[lawastooshort]

Nice stuff, Gandalf.

I second the totals request (possibly with a percentage of fraction/6 - the quickest way to show how it differed from the expected 1/6), and am not that interested in a graph.

So the main thing that went wrong for Rav in our match was 4/10 catches being 1, and rolling only 55% of the expected 6s.

[Gandalf]

Re: totals. I actually have them coded in (ie I have worked out the totals) but haven't put them in the final thing yet, due to fatigue... I thought this would take 1 evening, and it took 3 so I was just desperate to get a "good enough" version out. So they shall be along soon. I'm sure I can do something re: percentages. I like seeing the table like this though, so you can see where the overall numbers hide important information, eg if I roll 50 2's out of 50 bonehead rolls that's actually really good luck even though my average roll was a lowly 2. What I want eventually is a truer stats-based set of numbers, once I am happy that I'm correctly storing the required roll. eg Maybe Rav rolling 55% of expected 6's wasn't an issue because (fictional example) he rolled 200% of the expected 5's and none of his rolls required a 6.

So, no graphs, this is only a stop gap set of data. Eventually, though, yeah definitely.

It's easily done in Excel

One of the benefits of a nice simple table like I've done is precisely this

New question... if I want to show how lucky you've been, let's say, you were luckier than 80% of people would've been. How would this best be presented, if I wanted to boil it down to one figure?
a) 80%
b) 0.8
c) 20%
d) 0.2
e) Something else

My personal preference would be b). This is similar to what BBM does, except a properly-calculated zero to 1 scale is more meaningful than BBM's number. eg Bonehead Chance : 0.855 Dodge Change : 0.42 Pick-up Ball chance : 0.12... I think it's fairly clear to the user what these numbers mean in context with each other?

[Idanmel]

How about 1.8 for someone who is 80% luckier than average?
I instinctively see 0.8 as someone who is 20% unluckier than average.

[Gandalf]

But... the fact that all the numbers will be less than 1 is a bit of a giveaway surely?

[Idanmel]

Maybe put a plus sign, like this: +0.855

[Raveen]

When I was fiddling this afternoon I came up with the following: ({Average d6 roll}-3.5)/2.5 expressed as a percentage.

What that shows is the percent deviation from average for a d6. So if you roll all 6's then it's 100%, if you roll all 1's then it's -100%. if you roll 3's and 4's then you get 0%.

Another request, can we get the raw rolls for block dice please? With what you've got I can see the chosen result but not the other options. I'd like to have the plain dice.

Finally there's an oddity regarding the 2D6 rolls. I think the rolls are being attributed to the wrong team (this may match the letter of who rolls what in the rules). Worth checking. i'll endeavor to take a look.

Oh and table formatting isn't quite right on the block dice section of http://gandalfgames.net/bbdb/matchdetai ... tch_id=116 as La didn't get 15 3D blocks, that should be the total.

[Gandalf]

When I was fiddling this afternoon I came up with the following: ({Average d6 roll}-3.5)/2.5 expressed as a percentage.

What that shows is the percent deviation from average for a d6. So if you roll all 6's then it's 100%, if you roll all 1's then it's -100%. if you roll 3's and 4's then you get 0%.

That's nice.

Another request, can we get the raw rolls for block dice please? With what you've got I can see the chosen result but not the other options. I'd like to have the plain dice.

I'm trying to provide useful information to let you analyse how lucky (or otherwise) you are. The raw block dice play no part in that, only the "best" dice per roll does. You can roll 50% red skulls and still knock over everyone you try to hit. So no, I will not assist in providing data enabling poorly-reasoned conclusions to be drawn. Sorry for the strong words but I'm quite passionate about doing maths right.

Finally there's an oddity regarding the 2D6 rolls. I think the rolls are being attributed to the wrong team (this may match the letter of who rolls what in the rules). Worth checking. i'll endeavor to take a look.

If there is a problem it doesn't affect all rolls. The matches I've looked at looked vaguely right. But let me know what you find.

Oh and table formatting isn't quite right on the block dice section of http://gandalfgames.net/bbdb/matchdetai ... tch_id=116 as La didn't get 15 3D blocks, that should be the total.

Thanks, will look into it at some point.

edit : the tables have improved presentation now. Not saying it's *fantastic*, but it is better.

[Raveen]

That's nice.

Ignoring the patronising tone of that comment, I'll explain why I posted. I'm not a mathematician and neither will most of your audience be. For me the +-100% is a clear and comprehensible figure for comparison with the other team or other players. It doesn't matter what calculation lies behind it for it to be a useful presentation method.

I'm trying to provide useful information to let you analyse how lucky (or otherwise) you are. The raw block dice play no part in that, only the "best" dice per roll does. You can roll 50% red skulls and still knock over everyone you try to hit. So no, I will not assist in providing data enabling poorly-reasoned conclusions to be drawn. Sorry for the strong words but I'm quite passionate about doing maths right.

Please define best as context is massively important to the best order of rolls. If it's purely knock down then your analysis is flawed.

edit : the tables have improved presentation now. Not saying it's *fantastic*, but it is better.

Cool, I'll take a look. One other titbit, in the Spiders/Cheesers game (assuming the table formatting is correct) there was a 12 rolled on the 2D6 table so it's definitely not reporting 2D6-1. Could the 1 be caused by a modifier? A very peculiar implementation of Mighty Blow perhaps? I can't think how it could end up with a 1 but there aren't many skills or events that can affect 2D6 rolls.

[Gandalf]

Ignoring the patronising tone of that comment, I'll explain why I posted. I'm not a mathematician and neither will most of your audience be. For me the +-100% is a clear and comprehensible figure for comparison with the other team or other players. It doesn't matter what calculation lies behind it for it to be a useful presentation method.

Patronization comes from the POV of "read what I have already posted in this thread" - that this is a stop-gap, proper awesome accurate stats to come later on (and yes they will be easily digestible), so I'm not investing time & creating much from these numbers.

Please define best as context is massively important to the best order of rolls. If it's purely knock down then your analysis is flawed.

I think we had this discussion elsewhere. However 90%+ of blocks do come down to knocked over = success, turnover = failure, other = neither. Although imperfect, it's a better basis than total block dice rolled (see example earlier in thread).

Re: 2D6, yes, I've seen some 12's around too, maybe the issue is localised to chainsaws. I'll investigate that avenue sometime soon. If anyone sees any 1's on 2D6's elsewhere, let me know.

[Raveen]

Any evidence to support your figure of 90%?

I strongly suspect it varies wildly from race to race and coach to coach. I doubt that my Elves are that interested in knocking people down for example. Often a push will be as much of a success as a pow and more so than a BD.

You're right on Skulls though, they're a clear fail although they clearly aren't the only way to fail.

[DaigaroOgami]

Yeah - can I ask a probably obvious question - but do the dice rolled tables listed on under the team names respond to what that team rolled or what happened to that team - intuition tells me it should be the former but from looking at some Ducklings games it appears to be the latter.

And I know you said no graphs at this time; I totally understand and all is cool with me (I think fair play to you for doing the work while we all try and fill the horrid gap BBM has left us... the French Git).

However, in the long term I totally call for graphs because I (and probably a lot of people who will end up using this - Percy reckon's its going to be having a very bright future...) find looking at a screen of numbers quite tedious and uninspiring; pretty graphs and graphics is the future I look forward to!!!!!