Estimates of the numerical strength of armies are of doubtful validity in ancient historical literature and indeed in accounts of modern warfare even as late as the eighteenth century. Commanders-in-chief and official chroniclers have combined to obscure the truth by underestimating the number of their own troops and overestimating that of the enemy's. Greek historians from Thucydides onwards achieved admirable precision and objectivity in estimating the number of troops on their side, which is in striking contrast to the utterly unreliable figures given in oriental literature, but even they are of little value whenever there is reference to the Persian armies, whose size is wildly exaggerated even by the trustworthy and experienced eyewitnesses of Alexander's anabasis. The modest figures quoted for Greek armies during Classical times increase abruptly with Alexander's expedition and become still larger in the armies of the Diadochs, but this is explained by the dimensions of the undertakings, the economic resources then available, and the system of recruitment. Alexander won the day at Gaugamela with 40,000 infantry and 7,000 cavalry (Arr. Anab. 3.12.5), but by the time of the battle of Ipsus a record had been set with approximately 80,000 on either side (Plut. Demetr. 28.3), and, as will be seen later, the tendency towards ever greater numbers of soldiers was more prevalent in the Seleucid armies than among their contemporaries. Nevertheless, since the figures quoted for the Seleucid campaigns are not always derived from first-class sources, they require careful examination.