Wars can be declared while snorlax is up

[Kid Winchell]

But that doesn't solve anything? 60 seconds later a ton of people would be back competing for the second slot. 60 seconds after that a bunch of people would be competing for the third slot - and maybe two of them would declare at the same instant and both get slots. 

 

I've only seen 4 defensive war slots happen twice ever. It's just crazy unlikely luck. It could still happen even with annoying delays and timers. It's not a problem unless people find methods of exploiting it in my opinion.

The extra slot happens, in practice, when people load the declare war page while a nation is in beige and then declare war exactly when the hour changes. If you apply a rate limit, then the people who got blocked out will have to click through all the buttons again, which some people will necessarily do faster than others. And if Sheepy chooses an arbitrary (or even random) delay that's not a round number, then people won't be able to watch the clock to time things perfectly any more, so the issue won't come up.

[Aisha Greyjoy]

just set the script to fire after the captchas have been solved (you can solve and wait)

I tried that, not the script part, the waiting part.  Had to resolve after a minute or two.

[Kid Winchell]

If you can regularly duplicate this issue, I'll set a random sleep timer on the war dec page that would solve this. But I don't think what you're suggesting is feasible.

Not feasible? More like trivially easy. And exploitable.

Edited October 2, 2016 by Roll Sheepy

[Dr Rush]

Not feasible? More like trivially easy. And exploitable.

Why stop at 4? Why not do 50 just to really make the point?

[Kid Winchell]

Why stop at 4? Why not do 50 just to really make the point?

It's technically feasible, but finding that many people who are all in the same range, sober (Jess ) and don't have their clocks 2 minutes fast (Wilhelm ) and then managing them all would be, in the words of my wife, a !@#$ing nightmare. It was annoying enough trying to do with just 6-8 people.

[Timmy]

It's technically feasible, but finding that many people who are all in the same range, sober (Jess ) and don't have their clocks 2 minutes fast (Wilhelm ) and then managing them all would be, in the words of my wife, a !@#$ nightmare. It was annoying enough trying to do with just 6-8 people.

 

I mean, we could try it if you wanted us to

[durmij]

Not feasible? More like trivially easy. And exploitable.

Reported for war slot filling

 

Kappa

[Dr Rush]

It's technically feasible, but finding that many people who are all in the same range, sober (Jess ) and don't have their clocks 2 minutes fast (Wilhelm ) and then managing them all would be, in the words of my wife, a !@#$ nightmare. It was annoying enough trying to do with just 6-8 people.

In other words you need Mensa.

 

Anyways it occours to me that if this works with getting extra defensive slots, does it work with getting extra offensive slots?

[Woot]

6-8 people online and in range at once to get one extra warslot 50% of the time using scripts doesn't seem that trivial. And it happens so incredibly rarely in normal circumstances that anybody who does it a couple times will be reported as a cheater, so it doesn't seem that exploitable. I still don't think it's worth having to attempt every contested war declaration 3 different times because of an annoying rate limiter or something like that. 

[Kid Winchell]

In other words you need Mensa.

 

Anyways it occours to me that if this works with getting extra defensive slots, does it work with getting extra offensive slots?

No, because the game prevents you from executing an action on a page if you've loaded any other pages after it (in a new window/tab). You'd have to declare the wars sequentially, which means the game would have plenty of time to process wars.

 

6-8 people online and in range at once to get one extra warslot 50% of the time using scripts doesn't seem that trivial. And it happens so incredibly rarely in normal circumstances that anybody who does it a couple times will be reported as a cheater, so it doesn't seem that exploitable. I still don't think it's worth having to attempt every contested war declaration 3 different times because of an annoying rate limiter or something like that. 

6-8 people isn't all that many compared to the activity levels that most alliances see during war time (especially during blitzes, which is when this would be widely used). And it's only rare at the moment because the technique has yet to be widely known. Our success rate was low because it was the first time we tried it. (Also some people were drunk, Jess .) Once people were familiar with the technique we did a much better job. And the nature of the script is so trivial that it can be reduced a single action in addition to the regular war declaration process. And the solution to this problem is simply adding a random delay to each request. Which you wouldn't even notice, because 95% of the time it takes a page to load is because of connection speed, not waiting for scripts to execute on a server.

[Ayayay]

Seems like an easier solution would be to have the wars recalculated when attacks are actually launched. 4 people could still declare on a single person, but the first one to actually try to attack gets their war slot dropped.

[Kid Winchell]

Seems like an easier solution would be to have the wars recalculated when attacks are actually launched. 4 people could still declare on a single person, but the first one to actually try to attack gets their war slot dropped.

Not really. That would require a fair bit of active development to create and test the new behavior, while introducing a random delay is just one line of code that can be added without having to change anything else.

[Kosonome]

I don't know how is Sheepy code, but if it's making a select to get wars counts with one SQL query, and inserting with another SQL query, you can do with one single SQL query, reducing the odds of happening.

 

INSERT INTO wars (column1, column2)
SELECT 'Value 1', 'Value 2'
FROM wars
WHERE (SELECT COUNT(*) FROM wars WHERE nation_id = xxx) < 4;
 

Another way of reducing the odds is making a transaction rollback.

 

1) Start transaction

2) insert war

3) count wars

4) if there is 4 wars, rollback transaction

 

Or both things. It's just a example, I never did myself that, maybe useful for Sheepy.

Edited October 2, 2016 by Koso

[Avruch]

Ok, so the exploit has been proven. So take war decs that come in, lock further decs after a count exceeds 3, and then recount and eliminate any after 3 before allowing any attacks. 

[Kid Winchell]

I don't know how is Sheepy code, but if it's making a select to get wars counts with one SQL query, and inserting with another SQL query, you can do with one single SQL query, reducing the odds of happening.

Refactoring code to speed it up might solve the issue, but we can't say how much without knowing what it looks like already. And Sheepy's a lazy dev who probably won't want to do that.

[Buorhann]

Damn.  Ghux got to it before we could in Mensa.  Why'd you have to ruin it, Ghux?  We were going to exploit it hard~

 

 

 

I kid, but am I?

[LordRahl2]

He said the people had to be sober.  MENSA is not &#33;@#&#036;ing sober enough to do this.