# Squeezing scientific paper to fit within page limits

I am submitting a paper to a conference that has a limit of 10 pages. I have edited and edited the text, but the paper still fills 11 pages, and I am reluctant to remove anything. What tricks can I use to make the paper fit in 10 pages?

I know about \vspace{-1mm}, but are other and better tricks? How can I squeeze a little bit between lines, around the elements produced by \maketitle, around headlines, etc. in a way that is consistent and does not look too ugly? Is it possible to reduce the font size from 11 points to, say, 10.9?

(This should be community wiki, but I don't have enough rep yet.)

-
You can always make you own questions CW. Edit the question and check the CW box. – Loop Space Jul 30 '10 at 18:54
Do we have "don't-do-this" tag?.. – Grigory M Jul 30 '10 at 18:59
Why should this be community wiki? – Kevin Vermeer Jul 30 '10 at 19:44
Are you using a specific Latex class that is required by the conference (in which case you shouldn't use any such tricks...) or are you voluntarily using something like article (in which case there are several possibilities for saving some space, e.g., by using a more compact document style)? – Jukka Suomela Jul 30 '10 at 19:49
@Seamus: For many conferences, your suggestion will lead to an automatic reject. – Ben Mar 24 '11 at 8:14

Keep in mind that by doing this you will be doing one or more of the following: annoying the reviewers, making your paper less readable, shooting yourself in the foot. I've been know to play these tricks when I need to reduce it by a line or two, but 10% is pushing the limits. Its even worse if its a grant proposal. It may just be rejected out of hand, with no recourse. Part of being a good researcher is knowing how to write concisely, and knowing what can be left out.

-
+1! Editing the paper to make it more concise and clear (=better) is always preferable to editing the layout to make it smaller (=worse) – Kevin Vermeer Jul 30 '10 at 19:49
Good point, I agree with you however this does not answer the question. – sixtyfootersdude Oct 15 '10 at 9:34
-1, this doesn't answer the question. I also think its obnoxious to assume that the article can be shortened, without even having seen the article. – goblin Aug 31 '15 at 7:22
-1, does not answer the question. – berkay Jan 6 at 19:52

Under the constraints that you should not change any of the specifications that the committee lays out, the following work well, and can save you upto a page or more in a 10-11 page document.

• Use a compact bibliography style like abbrv instead of the defaults
• Preface the bibliography with a \small
• Tables of results can often be typeset one size smaller without any visual impairment. In fact they often look better that way, because they stand out from the text in any case. Using \small as above does the trick.
• As mentioned above, try to find and reword paragraphs that have dangling last lines with 2-3 words in them. This can make a bigger difference than you think.
• Use the package wrapfig to wrap figures around text, rather than having them set off and wasting space around them. This is very effective for small illustrative figures - not such a good idea for figures of results though.
• Conference style files (ACM, can you hear me?) often waste a lot of space around subheaders like \subsection and \subsubsection. Often, you can get around this by replacing organizational elements by \paragraph*. While this might kill the numbering, it is often reasonable, and saves a lot of space.

Update: Another trick I forgot to mention is using the microtype package. It's very effective at doing subtle shrinkage that undetectable to a non-font expert.

-
Do not \usepackage{times}. Read this. – TH. Aug 28 '10 at 13:46
Prefacing the bibliography with \small breaks the style decided upon by the committee. It is very commonly used though. – Ben Mar 24 '11 at 8:17
You should use TeX Gyre Termes via \usepackage[T1]{fontenc} \usepackage{tgtermes} instead of \usepackage{times} – Canageek Nov 14 '11 at 22:29
For Times, I use newtxtext and newtxmath. – Alexey B. Mar 9 '15 at 23:29
Can you link or elaborate regarding how to use microtype to achieve shrinkage? – einpoklum Apr 8 '15 at 7:11

I generally don't like to fiddle with fonts, margin or spacing because it can look very ugly, many journals and conferences explicitly ask you not to do this, and it just feels kind of cheating.

Having said that, I do have a cheat which I usually apply when I need to squeeze a few extra lines in the paper (specially if I just have little time before the deadline): Go looking for paragraphs that end with a few words (maybe up to three or four) on its last line, and try to edit those paragraphs first. See if you can somehow reword things or cut some adjective here or there so that you can kill those maybe three words and get a whole extra line for your paper!

Some other more obvious tricks: if you have displayed formulas, consider if you can have them inlined with the text instead, and if you have figures consider scaling them down a bit.

-
+1 for removing widows at the end of paragraphs. – onewhaleid Jan 19 '15 at 4:51

Check out this page on Squeezing Space in LaTeX, it contains many useful tips ranging from using standard stuff like \linespread{0.9} to using packages that allow more customization like paralist or caption.

-
paralist is great ! – Suresh Aug 3 '10 at 17:17

For 'squeezing' material of any kind, the savetrees package is often a good place to start. It has a number of levels of squeezing, but even if you don't use the package you can pick up some tips from there. (I'm also agree with the other answers about worrying about the wisdom of looking for this much of a squeeze.)

-

Find a long paragraph with only 1 or 2 words on the last line, and say \looseness=-1 immediately after the last word of the paragraph, to ask tex to try to reset that paragraph a line shorter.

-

You might want to consider using a document class that is more space-efficient than article. I have fairly often used amsart – I think it looks reasonably good, and it doesn't waste as much space as article.

-
Changing the document class leads to automatic rejection for many conferences. – Ben Mar 24 '11 at 8:18

Sometimes I have managed to gain upto 5-6 lines by forcing TeX to break, a page at a point that it considers ugly. I do this by putting \goodbreak\noindent at the end of an existing line break in a paragraph that is close to a page boundary. Normally, this is enough to force a page break at that spot.

-
Huh, I don't think I've used \goodbreak before. Nice command to keep in mind. – Will Robertson Nov 22 '10 at 4:44

I described several tools for this at my blog a few months ago.

Update: I have copied and pasted my post below.

Make your text block as big as possible. The simplest way to do that is using the geometry package:

\usepackage[text={16cm,24cm}]{geometry}

Use a compact font such as Times Roman:

\usepackage{mathptmx}

Remove the spacing between paragraphs and have a small paragraph indentation

\setlength{\parskip}{0cm}
\setlength{\parindent}{1em}

\usepackage[compact]{titlesec}
\titlespacing{\section}{0pt}{2ex}{1ex}
\titlespacing{\subsection}{0pt}{1ex}{0ex}
\titlespacing{\subsubsection}{0pt}{0.5ex}{0ex}

Beware of enumerated and itemized lists. Instead, replace them with compact lists.

\usepackage{paralist}
\begin{compactitem}
\item ...
\end{compactitem}
\begin{compactenum}
\item ...
\end{compactenum}

If you are allowed, switching to double column can save heaps of space.

\usepackage{multicols}
\begin{multicols}{2}
...
\end{multicols}

If the rules say 12pt, you can usu ally get away with 11.5pt with out anyone noticing:

\begin{document}\fontsize{11.5}{14}\rm

When you get desperate, you can squeeze the inter-line spacing using

-
Could you briefly describe a few of those tools/tricks here? It's fine to include a link to your website but, whenever possible, we want answers to be hosted in the site. See meta.tex.stackexchange.com/questions/227/… – Juan A. Navarro Aug 3 '10 at 7:15

In the article class, I think the output of \maketitle is both ugly and wastes too much space in a short document. If you don't want to use another document class, a simple solution is to design your own replacement of \maketitle.

If you don't want to worry about creating your own \maketitle macro, you can simply omit commands such as \title and \maketitle and hard-code the title by using, e.g., center environment, appropriate font selection commands, and \vspace.

-

\section{Introduction}

This is both boring and redundant (a paper should always start with an introduction). If your introduction does not contain any subsections, just skip the \section command.

-
This was the single most prudent suggestion. No compromise on formatting involved, no trickery, just plain unadulterated common sense :) – Vaibhav Garg Feb 6 '14 at 6:49

I would 'concur' with a reply above but don't have the 'points' to do so-

I'd be wicked careful about this, especially when it comes to grant proposals- I've sat on more than one panel where a submission was returned to the author un-reviewed because of these shenanigans. I've also witnessed this with respect to a conference abstract even.

I realize this doesn't 'answer' the question, as is the goal / rewarded behavior here, but thought I'd just chime in on it none the less for those of you contemplating it.

-

When I'm reviewing conference papers, I don't mind limited margin/font/spacing abuse to get the paper under the limit; my favorite is to make the font size in the references smaller, or eliminate blank lines from the references. However, such abuse which affects the readability of the paper in any meaningful way would earn an immediate rejection recommendation from me, and a 9% reduction (if you really do have a full page to go) really is quite a lot to expect from this.

And for the camera-ready copy, such tactics are totally unacceptable if they're noticeable at all.

Really, the only correct answer is "edit it more". Having been there numerous times, I am 100% sure you can edit the paper down another page.

Perhaps you could say which conference you are submitting to. If it's an ACM conference, I might be able to give more advice.

-
If it's an ACM conference, it basically doesn't matter, it'll look ugly no matter what and their provided class is terrible. =\ – TH. Aug 28 '10 at 13:48

Is there a lot of math in your document? If you suspect that the blocks of displayed math are taking up too much space (and possibly interfering with natural column/page breaks), then you might want to try these tricks:

• Reducing the font size of only the math environment block

• Allowing breaks within the math environment block

To achieve this, you can define the following commands in the preamble:

\newcommand{\startcompact}[1]{\par\vspace{-0.75em}\begin{#1}%
\allowdisplaybreaks\ignorespaces}

\newcommand{\stopcompact}[1]{\end{#1}\ignorespaces}

and use them in your document like this:

\startcompact{small} % or footnotesize, scriptsize, etc.
\begin{align}
E &= mc^2\\
a^2 + b^2 &= c^2
\end{align}
\stopcompact{small}

If you want to allow breaks for all displayed math blocks, you can just put \allowdisplaybreaks in the preamble on its own; then to prevent breaks across individual lines, simply use \\* instead of \\.

This approach works quite well for the IEEE templates. (I would even argue that the compiled document looks better.)

-

Back in the day, people only used passive voice in formal papers, but now active voice is becoming more acceptable. Find a few sentences where you wrote in passive voice and rewrite them in active voice (this usually improves readability too). Also remember to never use a long word when a short one will do. Also, try shrinking any images or graphs by around 10% (but don't do this if it makes them unclear).

-

You can reduce the base font size in the document class statement

\documentclass[10pt,other options]{article}

which will make everything smaller. Your choices for that base font size are limited, though; I think it may only allow 10pt, 11pt, and 12pt. There are also various packages to change font sizes; have a look at relsize and anyfontsize, for example.

Another option would be to reduce the margin size by using a package like vmargin.

-
I think most conferences specify font sizes too. – ShreevatsaR Jul 30 '10 at 19:23
Even if the conference doesn't so specify, this likely wouldn't help the OP. The default size is 10pt, so changing from no specified base size to a specified one is either a non-op or going the wrong way. – vanden Jul 30 '10 at 21:00
Changing fonts sizes and margins is usually severely frowned upon in my experience. – Ben Jun 9 '11 at 16:06
In many conferences, guidelines include something like "papers with smaller page margins and font sizes than specified in the author instructions and set in the style files will also be treated as overlength". Take care if you decide to use this trick. – Elenaher Apr 16 '14 at 15:43

this is quite a common problem I guess. The first thing you should do is to make sure that the figure captions are \small. You can do this by adding

\usepackage[font=small]{caption}

I personally believe that small captions look better because they are easier to distinguish from the main text.

You can also try reducing the size of some very big equations. I usually use the following kind of code for this purpose:

$$\resizebox{.95\columnwidth}{!}{ whatever equation you want }.$$`

Some equations look even better with this trick, but do not exaggerate ;-).

You can reduce the space around figures with \vspace{-1mm} as you suggest.

Finally, if I am really desperate and 20 minutes from a deadline, I just add something like

\def\baselinestretch{0.99}

in the preamble. This can give quite surprising results, in terms of space reduction, because it can induce a general rearrangement of figures and text even for values very close to 1. However I agree with KeithB that this (as other similar tricks) reduces the quality of your paper. As a rule of thumb, I never go below 0.97 and I remove this in the final version if the paper gets accepted.

In general you should try to reduce the content first. This can ALWAYS be done and is definitely the best solution for the reviewer.

I just released that this question is very old. Well just in case my answer can be useful to somebody.

-
Most of these suggestions may go against the journal's setup, apart from "reducing the [actual] content": (1) Captions usually have a specific style, so setting them using \small would most likely be reverted; (2) \resizebox for equations also requires the occasional \displaystyle otherwise everything is formatted in \textstyle... and everything will shrink; (3) Better be consistent with spacing around floats, for which you can see Remove space after figure and before text. My two cent's worth. – Werner Jan 19 '15 at 5:11
You don't get to typeset the journal version. If accepted you will have all these tricks removed and you will pay extra page charges so this will be useless and in fact annoying for the reviewers. – percusse Jan 19 '15 at 7:43