# Flush matrix entries to Left side

So I have the following codes

\begin{dmath}
P_1=\left\{\begin{matrix}
{P_E} + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\
- KI(0,T,S,{B_{11}},{b_{11}}, - 1,r)\\
+ SI(0,T,S,{B_{11}},{b_{11}},1,\delta )\;\text{if }\;S > {B_{11}}\\
K - S\quad \text{if } \;S \le {B_{11}}
\end{matrix}\right.
\end{dmath}

\begin{dmath}
P_2=\left\{\begin{matrix}
{P_E} + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\
- KI(0,T/2,S,{B_{21}},{b_{21}}, - 1,r)\\
+ SI(0,T/2,S,{B_{21}},{b_{21}},1,\delta )\\
- KI(T/2,T,S,{B_{21}},{b_{21}}, - 1,r)\\
+ SI(T/2,T,S,{B_{21}},{b_{21}},1,\delta )\quad \text{if }\;S > {B_{22}}\\
K - S\quad \text{if }\;S \le {B_{22}}.
\end{matrix}\right.
\end{dmath}

\begin{dmath}
P_3=\left\{\begin{matrix}
{P_E} + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\
- KI(0,T/3,S,{B_{33}},{b_3}, - 1,r)\\
+ SI(0,T/3,S,{B_{33}},{b_3},1,\delta )\\
- KI(T/3,2T/2,S,{B_{32}},{b_{32}}, - 1,r)\\
+ SI(T/3,2T/3,S,{B_{32}},{b_{32}},1,\delta )\\
- KI(2T/3,T,S,{B_{31}},{b_{31}}, - 1,r)\\
+ SI(2T/3,T,S,{B_{31}},{b_{31}},1,\delta )\quad \text{if }\;S > {B_{33}}\\
K - S\quad \text{if }\;S \le {B_{33}}.\\
\end{matrix}\right.
\end{dmath}


Unfortunetly this produces the following picture

I want it to look more like this

• Welcome to TeX - LaTeX! Have you tried to use a cases environment, possibly with an aligned inside? – Andrew Swann May 22 '17 at 16:42
• no I have not, first time ive heard of it – asdf May 22 '17 at 16:45

I would use the cases environment from the amsmath package. I don't know where your matrix and dmath come from.

I think the extra spacing you experience is due to the \left\{.

Example:

\documentclass{article}
\usepackage{amsmath}
\begin{document}
\noindent
$$P_{1}= \begin{cases} P_E + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\ {}- KI(0,T,S,{B_{11}},{b_{11}}, - 1,r)\\ {}+ SI(0,T,S,{B_{11}},{b_{11}},1,\delta )\quad\text{if}\; S > {B_{11}}\\ K - S\quad \text{if}\; S \le {B_{11}}. \end{cases}$$

$$P_{2}= \begin{cases} P_E + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\ {}- KI(0,T/2,S,{B_{21}},{b_{21}}, - 1,r)\\ {}+ SI(0,T/2,S,{B_{21}},{b_{21}},1,\delta )\\ {}- KI(T/2,T,S,{B_{21}},{b_{21}}, - 1,r)\\ {}+ SI(T/2,T,S,{B_{21}},{b_{21}},1,\delta )\quad \text{if}\; S > {B_{22}}\\ K - S\quad \text{if}\; S \le {B_{22}}. \end{cases}$$

$$P_{3}= \begin{cases} P_E + K(1 - {e^{ - rT}}) - S(1 - {e^{ - \delta T}})\\ {}- KI(0,T/3,S,{B_{33}},{b_3}, - 1,r)\\ {}+ SI(0,T/3,S,{B_{33}},{b_3},1,\delta )\\ {}- KI(T/3,2T/2,S,{B_{32}},{b_{32}}, - 1,r)\\ {}+ SI(T/3,2T/3,S,{B_{32}},{b_{32}},1,\delta )\\ {}- KI(2T/3,T,S,{B_{31}},{b_{31}}, - 1,r)\\ {}+ SI(2T/3,T,S,{B_{31}},{b_{31}},1,\delta )\quad \text{if}\; S > {B_{33}}\\ K - S\quad \text{if}\; S \le {B_{33}}. \end{cases}$$
\end{document}


• to be precise they are not aligned p1, p2 , p3 but anyway thanx – asdf May 22 '17 at 16:45
• nvm I solved it. – asdf May 22 '17 at 17:23
• You may want to prefix the - and + symbols at the beginnings of lines with {}, to make TeX interpret them as binary rather than as unary operators. – Mico May 22 '17 at 18:14

If I understand correctly your intent, you are after something like this:

The code:

% My standard header for TeX.SX answers:
\documentclass[a4paper]{article} % To avoid confusion, let us explicitly
% declare the paper format.

\usepackage[T1]{fontenc}         % Not always necessary, but recommended.
% End of standard header.  What follows pertains to the problem at hand.

\usepackage{mathtools}

\begin{document}

\begin{align}
P_{1} &=
\begin{cases*}
\begin{aligned}[b]
P_{E} &+ K(1 - e^{-rT}) - S(1 - e^{-\delta T})\\
& - KI(0,T,S,B_{11},b_{11},-1,r)\\
& + SI(0,T,S,B_{11},b_{11},1,\delta )
\end{aligned}
& if $S > B_{11}$, \\
K - S & if $S \le B_{11}$.
\end{cases*}
\\
P_{2} &=
\begin{cases*}
\begin{aligned}[b]
P_{E} &+ K(1 - e^{-rT}) - S(1 - e^{-\delta T})\\
& - KI(0,T/2,S,B_{21},b_{21},-1,r)\\
& + SI(0,T/2,S,B_{21},b_{21},1,\delta )\\
& - KI(T/2,T,S,B_{21},b_{21},-1,r)\\
& + SI(T/2,T,S,B_{21},b_{21},1,\delta )
\end{aligned}
& if $S > B_{22}$, \\
K - S & if $S \le B_{22}$.
\end{cases*}
\\
P_{3} &=
\begin{cases*}
\begin{aligned}
P_{E} &+ K(1 - e^{-rT}) - S(1 - e^{-\delta T})\\
& - KI(0,T/3,S,B_{33},b_{3},-1,r)\\
& + SI(0,T/3,S,B_{33},b_{3},1,\delta )\\
& - KI(T/3,2T/2,S,B_{32},b_{32},-1,r)\\
& + SI(T/3,2T/3,S,B_{32},b_{32},1,\delta )\\
& - KI(2T/3,T,S,B_{31},b_{31},-1,r)\\
& + SI(2T/3,T,S,B_{31},b_{31},1,\delta )
\end{aligned}
& if $S > B_{33}$, \\
K - S & if $S \le B_{33}$.
\end{cases*}
\end{align}
Text after the equations.

\end{document}


Points to note:

• The mathtools package (which automatically loads amsmath too) provides a cases* enviroment that typesets the second column in text—as opposed to math—mode. Other choices are also available: see p. 18 of the manual of the mathtools package.

• I’ve used an aligned environment, with optional argument [b], to wrap long formulas that (I repeat, if I understand correctly) belong to the same logical branch.

• I’ve also removed a lot of superfluous braces.

• As a “feature” (;-) , the last equation shows, by way of example, what purpose the [b] argument serves…