Hey, I wanted to start new series about hacking. I think that this part of programming contests is really underrated, but can be helpful for both people solving problems (you should know what tricky cases you need to consider) and people writing problems (you should know how to write nice tests covering everything).

So in this series, which I will try to continue, I will talk over hacks at Codeforces Rounds. I will try to do some statistics and take a look at most common mistakes in passing contests.

It is my first article related to hacking (and first at all :)), so feel free to suggest something, correct me on some errors or mistakes (my English still can be much better :)) and tell about your hacks. Thanks in advance!

So let's move to Codeforces Round 262 (Div. 2).

Update: added country wise statistics (at the end).

Update: added chart with hacks over time and first hacks.

Update: the handles are now full of colours, with many thanks to awesome accidentallygivenfuck!

Update: added section "biggest gap" (between rating of the hacker and defender).

Stats

Really nice graph made by soon (thank him here):

And my graph showing hacks over time:

Remarks

It looks like problem B was the most grateful for hackers (above 600 successful hacks and it was nearly of all tries!). But still, every solution has at least one successful hack.

460B - Little Dima and Equation

This problem had a lot of successful hacks! The most frequent mistakes were:

• printing solutions that are negative or bigger than 10⁹
• forgetting about using 64-bit numbers (sometimes numbers can be too big for 32-bit)
• wrong upper bound for sum of numbers, it should be at least 81 (more than this, like 100, was ok)

And there was pretty wrong solution which can pass the pretests: checking all numbers to some number fitting into time limit (like 1000000).

460C - Present

There were two most frequent types of hacks. First one was just giving big n, m and w. In this case solutions with complexity O(nm) described in tutorial was failing. Second one was using the fact that the result can be bigger than 10⁹. Many people using binary search just forgot about it and set upper bound to just 10⁹. Simple hack:

1 10000 1
1000000000

The result should be 10⁹ + 10⁴ in many programs it was around 10⁹.

460D - Little Victor and Set

All hacks were just big x and y and k = 3. It was destroying naive approach in such case of k.

460E - Roland and Rose

The one and only hack for problem E was this one, the solution reached time limit.

Fast hackers

The one and only hack for E was done in the very last minute of contest. Congratulation 4erneff!

Best hackers

Best rooms

Countries

Country	#hacks	Hackers (with more than one successful hack)
China	133	Nero [9], skyxuan [7], hnshhslsh [7], Memset137 [7], rabbitlover [6], aswmtjdsj [6], vengarth [5], Devasundaram [5], nhywieza [4], lveternal [4], DanielChau [4], Napoleon_cn [4], FinalTheory [4], I_Love_Balabala [4], Dwayne [4], MADAO_G [4], xwind [3], lin11 [3], bolyeria [3], Just_ck [3], TaoSama [3], CST14-Lee [3], Atlantis67 [3], qingping95 [2], polossk [2], dnvtmf [2], wyxourlove [2], Heart_Blue [2], Harry.Guo2012 [2], wwunxc959 [2], 0sFr [2]
India	88	pk. [8], mohit.829 [7], prathyu [6], codefor6768 [6], n1tz53 [5], yellow1 [4], shivatejesh [4], sanjeev1779 [4], Revs [4], win_ay39 [3], shubh_jain [3], pranet [3], boygenius [3], SameerGulati [3], osank [2], mmaks [2], meintoo [2], m17 [2], davinci648 [2]
Russia	66	linjek [10], SC_pRo_ION [6], Guliash [6], rureggaeton [5], Sehnsucht [5], qwerty787788 [4], Artem_kh [4], holycow [3], Slyusar [3], Kvark161 [3], worse [2], stalker506 [2], ItsLastDay [2], Androik [2], ABalobanov [2]
Ukraine	44	Programist [7], Dudka [7], chakred_namor [6], Antoniuk [5], Rebryk [3], Omelianenko [3], BigBag [3], mathhislife [2], konstantin.lex [2], fiter [2], aangairbender [2], T0RRES [2]
Bangladesh	40	Honour_00 [10], the_redback [8], Baridcse [5], ndatta [4], raihatneloy [3], FlaminRage [3], MKGAURAB [2], Binary_ToothLess [2]
Turkiye	36	yEmre [8], muratt [7], kalimm [7], ikbal [6], kursatbakis0 [3], ykaya [2], KayacanV [2]

You can see that worse had 2 successful hack attempts (and 22 unsuccessful :)). And let's see what will happen if we take all successful hacks and divide by number of people who tried to hack (with at least one hack attempt):

Country	#hacks / #hackers	Hackers (with at least one successful hack)
Colombia	6.00	jhonber [6]
Turkiye	4.50	yEmre [8], muratt [7], kalimm [7], ikbal [6], kursatbakis0 [3], ykaya [2], KayacanV [2], WorstCoder [1]
Serbia	4.00	VladaMG98 [11], allllekssssa [1]
Cyprus	4.00	michael10024 [4]
Belgium	4.00	jorik [4]

Biggest gap

In this section I took all successful hacks and looked at difference in rating between hacker and defender. There were 18 "new" hackers (it was their first contest), so I simply miss them from obvious reasons.

Thanks for your feedback!