I Feel Like The Dumbest Person On Fibs

[snowjake]

I play on Parlor Play. I've tried to but can't figure out how to properly use JavaFibs. I can't make it do what I want to. I lose games constantly (I'm talking about my connectioon) and have 35 or so saved games yet I'm not a dropper.

I've read the post explaining the new function of repbot yet I don't have the slightest idea what it al means. I had a lot of vouchers but theu won't count at all anymore, right? Only my 35 saved games will be seen by strangers and they'll think I'm some sort of dropper, right?

BTW, as I said I play on Parlor Play. I can't even find where the Repbot went for the last 2 days.

[don]

You are far from the dumbest snowjake.

Part of your problem is that you are trying to use a GUI, JavaFIBS I think, and haven't quite gotten familiar with it.  The other part is that your GUI and MOST GUI's, don't really interface to FIBS in that they don't provide all the information.

Whatever you are using, you need to read some FIBS' help files to get up to real speed.  Typing "help" at your command line, whatever windoze it is in, will give you a clue.

Also, you need to realize what the kind of rep you are asking about involves.  Once you have read the FIBS help files, you will realize that completing games is important.  One way to up your chances of completing games is to do a who on the player, look at how many matches he's played, his ratings score.  That's clue one to whether you will get dropped or not.

Then you can look at repbot ratings, but the only objective info repbot puts out is number of saved matches.  It's an excellent indicator of the likelihood of your opponent to continue a match.

The vouches and complaints on repbot can be and are manipulated.

If you want some raw data for a person's reputation, data most FIBSters aren't aware of, ... :

    FIBSaved

Using available FIBS' info, I can see that Snowjake is:

Information about Snowjake:
 Last login:  Wednesday, July 19 21:11 PDT from ns1.livepenguin.com
 Last logout: Wednesday, July 19 21:13 PDT
 Not logged in right now.
 Rating: 1583.53 Experience: 958
 No email address.

and that Snowjake has saved matches with:

Bloo Cointoss Ernst_O FlyByNight Ged Hizli JeffK MadGreek RayBan RaysDaddy Studley VictorJ albatros bobmont brneyecute chatless delta demiurgos drick dsnine janitorb javier_uno joshuah jpb littlesu mmh natsy nicky playbackgammon prairieclimber rickelfox rradarr shadowcaster stonedcow sukhy tcassin thanos waggis

Anyway, a good reputation usually doesn't have so many saved matches with so low a number of games played.

--
don

[diane]

I have known that list of saved be out by as many as 5 games - so I have stopped using it so much - somehow repbot is finding more accurate numbers - so I would recommend taking the numbers from repbot.  No one 'knows' where repobot is when he goes offline - repbot is hosted by a player much like yourself - if he goes on holiday and his system goes down - he isnt there to put it back on - and we lose repbot for a day or two.  At that time - everyone takes a bit more care who they play with - I dont think I would worry about playing you, snowjake  

When repbot is back tell repbot 'toggle alert' as a priority - this means it will tell you when someone you have a saved game with logs in - it helped me get my saved games down from 30-40 to less than 5 - back when I had dial up and got logged off every two hours.

Hope this helps - and dont worry - we arent all coders - and a lot of this stuff isnt obvious - just take it bit by bit - you will find you learn more as you go along  

[lewscannon]

No one 'knows' where repbot is when he goes offline

I once saw repbot hanging out with Mookie, trying to get past backstage security at a Jewel concert, claiming that there was an emergency shoe repair that only he could perform.

[Slipshod]

You're not dumb Jake, you are just not gay enough.  Log on thursdays and you will see how much better you are with the "client"; you'll have that puppy smiling in no time.  

[Dipper]

I have known that list of saved be out by as many as 5 games - so I have stopped using it so much - somehow repbot is finding more accurate numbers -

That's encouraging to hear, Diane. How well-known is that?

I'm 3 days old on FIBS. I've played 17 matches. I've never dropped a match, by intention or accident. Yet that list shows *7* dropped matches -- listing opponents I had perfectly good and complete matches with.

It's a bit disheartening.

[burper]

Here is another idea for generating data that might prove useful:
Consider the full list of "started matches of the day" (http://fibs.com/savedgames/list.html) over time.

Process the list(s) in various ways to derive more information than a single days' list can provide.

If A plays B, and finishes, we should see A on Bs' list and B on As' list on the day they play, and not on the next. If we see either of these on future lists, we can tell that they have played another match, and in fact can count how many times A and B play. One could infer a recommendation of A from B and B from A, the strength of which would be proposional to the count.

If we see B on As' list for consecutive days, we know either that:
-they have a saved match
-they started a match on each of those days
Departing from strict objectivity, one *might* choose to believe the first option above is somewhat more likely than the other. If they see multiple such cases, it may mean something more than an individual case.

One could calculate the average age of saved matches on a players list.

Add the knowledge of how long matches stay on the list before expiring, and one could count the number of times a player *may* have allowed matches to expire during the last month.

If you think about it more, you could come up with other interesting statistics that could be boiled down, all without have to connect to FIBS as a bot. "Interesting" is in the eye of the beholder, so one could customize the processing of the stored raw data into something they believe is the most meaningful to them, or not process it at all and look at the overall data for the user (which comes down to the same thing reallly, just using doing the match in your head).

Now add the ability to be logged into FIBS, and to have access to RepBots' database, and there are far more possibilities.

Storage isn't that much of a problem, as I think all the lists generated from the beginning of FIBS (if they existed) would still fit in 1Gb, but processing time would be, so some accumulated format of data would be needed. XSD anyone?

RepBot could be improved and/or replaced, in ways that even don wouldn't be able to argue with, but noone is really interested that: people have too good a time whining about how it is not-objective and how it is being manipulated etc... These are the same people that have not yet realized that FIBS is an ongoing psychological experiment. Oops...

So I don't expect any serious responses to this, but what follows is a quick first pass that I am not entirely happy with, just in case. Note to non-coders: there is no excuse for not helping out, as you could ponder and discuss other objective data that can be generated from the "savedgames" list, and what interesting (and subjective) things about a user might be inferred, e.g. how many different players someone starts matches with during some online period, combined/compared with what other data, etc...

Code Select Expand


<?xml version="1.0"?>

<!--
XML schema for FIBS accumulated saved matches
The idea, is that this accumulation adds value to daily lists, e.g.
"age" of a player (days with account on fibs) could be compared to
experience, numbers of saved matches older than some filter value
might give a better indication of dropping potential,  large number
of matches against a particular player could imply a RepBot-style
"vouch", etc...
-->

<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
   targetNamespace="http://www.fibs.com"
   xmlns="http://www.fibs.com"
   elementFormDefault="qualified">

<!-- legal FIBS usernames, and <kdc>  :)  -->
<xs:simpleType name="fibsUsername">
   <xs:restriction base="xs:token">
<xs:pattern value="[a-zA-Z_]+"/>
<xs:maxLength value="20"/>
   </xs:restriction>
</xs:simpleType>

<!-- saved match info, with opponent and date first seen -->
<xs:complexType name="savedMatch">
   <xs:sequence>
<xs:element name="op" type="fibsUsername"/>
<xs:element name="da" type="xs:date"/>
<xs:any minOccurs="0" maxOccurs="unbounded"/>
   </xs:sequence>
</xs:complexType>

<!-- accumulated count of matches between players -->
<xs:complexType name="lifetimeOpponent">
   <xs:sequence>
<xs:element name="na" type="fibsUsername"/>
<xs:element name="co" type="xs:unsignedInt"/>
<xs:any minOccurs="0" maxOccurs="unbounded"/>
   </xs:sequence>
</xs:complexType>

<!-- a user record, with name, date first seen,
saved matches and lifetime opponents -->
<xs:complexType name="userRecord">
   <xs:sequence>
<xs:element name="na" type="fibsUsername"/>
<xs:element name="st" type="xs:date"/>
<xs:element name="ma" type="savedMatch"
    minOccurs="0" maxOccurs="unbounded"/>
<xs:element name="op" type="lifetimeOpponent"
    minOccurs="0" maxOccurs="unbounded"/>
   </xs:sequence>
</xs:complexType>

<!--
the root element of the fasm schema, containing:
    -meta data
    -list of user records
-->
<xs:element name="fasm">
   <xs:complexType>
<xs:sequence>
    <xs:element name="title" type="xs:string"/>
    <xs:element name="description" type="xs:string"/>
    <xs:element name="started" type="xs:date"/>
    <xs:element name="updated" type="xs:date"/>
    <xs:element name="ur" type="userRecord"
 minOccurs="0" maxOccurs="unbounded"/>
</xs:sequence>
   </xs:complexType>
</xs:element>

</xs:schema>


[diane]

That's encouraging to hear, Diane. How well-known is that?

I'm 3 days old on FIBS. I've played 17 matches. I've never dropped a match, by intention or accident. Yet that list shows *7* dropped matches -- listing opponents I had perfectly good and complete matches with.

It's a bit disheartening.

Apparantly it is all in the blurb on the page - when I questioned Patti about this - her reply was

"Note that it may include games that were concluded during the last 24 hours, as well as umlimited matches and matches in progress."

In short, if the "oldmoves" and "oldboard" commands are still available for the game, it will be counted as a saved match even if the match has been completed.

So - it is very well known - by those who understand that    

[Dipper]

Aha -- thanks -- on the tally of saved games page. I was looking at the blurb-less raw list of saved games.

(Burper - I start to get your logic, then. But you'll forgive a new guy for not investing many brain cycles in any of this just yet...)