League changes? RESOLVED (BY DESIGN)

[Peter]

Well, I was looking for a better place to post this, but okay.

In have run the leagues to get some ideas to improve M3 to M4

In DB2.43   Multiply3 comes in as 5th if I run the league.
In DB2.43v Multiply3 comes in as 29th if I run the league.


What bot-changes have there been made between 2.43 and 2.43v.

I know something about body chances, where there any further chances made. If it is just body it will probably become better if I change body management. So where there too any poison or venow or anything changes.

[EricL]

There have been many changes adn bug fixes since 2.43 that could impact league standings.  Poison didn't work in 2.43 for example.  Look here for the complete list.

[Numsgil]

You're talking like 6 months of changes, so I can't list them all off the top of my head.  Check out the buddy drop thread, Eric's good about listing every change.  Read through them all and see if any of them suggest the problem.  It could also be a bug.  Hard to say this early in the process.

[EricL]

Thinng to look at:  virus insertion now random, tie changes w.r.t. .tiepres and tied bot tie numbers, poison works, tie sharing constraints, the 100 per cycle limit on the .mkvenom attack and the new conditional paradym.

[Peter]

One small thing in DB2.43v it isn't 29th but just plain the last one, well it beats the last one out of the league.

I have the idea it wasn't one change that coused this dropping. I have just checked DB2.43o.

The war for the place before the last.

In DB2.43 : Multiply3 wins
In DB2.43o:statistical draw
In DB2.43v:Martian Tank3 wins


Well, it is kind of funny, it was running in as 5th, now 6 months development later it would be standing last, and in the middle it would give a draw against MT3.

[Jez]

Hiya Peter,

The leagues have never been re-run, by that I mean the bots in the league only got their position because of the performance they gave at the time they were tested.
This means that sometimes the top bot has become really rubbish (I.e. it was exploiting a bug when it was tested that has since been fixed) or they have suddenly become much better (again because of changes made to DB since they were tested).

I did try re-running the leagues once but 2BH it took ages and the version of DB at that time wasn't up to the job and so the league is still being run in the way PY originally intended it.

Even if I ever get off my arse and get round to adding the latest batch of bots to the league I have no intention of trying a full rerun of the leagues again...  

[Peter]

Even if I ever get off my arse and get round to adding the latest batch of bots to the league I have no intention of trying a full rerun of the leagues again...  

No need to do it eather Jez, I like this way, a complete rerun is probably not going to change league standings everytime you do it.
The changes aren't that big eather(just all kicking mult a little down  ), in DB2.43 multiply has some trouble beating MT3, and in DB2.43v a little too much trouble  .

I am not very sure how new MT3 (martian tank3) is, I thought it was put in in the last update.
But my bet is that it hasn't got a proper virus-defence and becouse of it, it lost against speedy(vir). Besides that, it is a really good fighter and therefore if speedy wasn't there I bet it was coming in top10.