From 1268 until the Republic fell in 1797, Venice elected its Doge through ten rounds that alternated drawing lots with voting. Thirty members of the Great Council were drawn by lot, cut to nine by a second draw, and the group was then shrunk by lot and enlarged by election four more times until a college of 41 electors chose the Doge on at least 25 votes (Molinari, 2018, p. 2).
The point of the alternation was to make the outcome impossible to arrange in advance. A faction could buy or bully electors only if it knew who they would be, and every lottery round destroyed that knowledge. Two further rules attacked the same problem directly: no family could hold more than one place in any committee, and a candidate's relatives were barred from voting on him (Molinari, 2018, p. 2). Modelling the procedure as a contest between two factions, Molinari (2018, p. 16) finds that it gave a minority a better chance of winning than a straight vote would, while keeping that chance below the minority's share of the assembly, so the larger faction was not penalised.
The procedure ran in rooms that are still open to visitors. The Chamber of the Great Council in the Doge's Palace, 53 metres long and 25 metres wide, held the assembly the first thirty names were drawn from, and from 1532 the neighbouring Chamber of the Scrutinio was used for electoral procedures until that became its only function (Palazzo Ducale, 2026). What follows sets out the ten rounds, the reasoning behind the numbers, and what five centuries of results actually show.
How was the Doge of Venice elected?
The protocol had ten rounds. Nine of them produced the electoral college for the round that followed, and only the tenth elected the Doge. Rounds alternated by type. In a lot round, the next college was drawn at random from the current one. In an election round, the current college chose a larger group, and every name had to clear a minimum number of approvals (Mowbray and Gollmann, 2007, p. 2).
The sequence ran as follows. Thirty members of the Great Council were drawn by lot, then reduced by lot to nine. Those nine elected forty, each name needing seven approvals of nine. The forty were reduced by lot to twelve, who elected twenty-five on nine of twelve. The twenty-five were reduced by lot to nine, who elected forty-five on seven of nine. The forty-five were reduced by lot to eleven, who elected the final college of forty-one on nine of eleven. Those forty-one elected the Doge, who needed at least twenty-five of their votes (Molinari, 2018, p. 3).
Five draws and five ballots, in strict alternation. Voting was by approval: electors marked each proposed name as approved or not, with no limit on how many they approved, and if too few names cleared the threshold the college simply balloted again (Mowbray and Gollmann, 2007, p. 3).
Why was the Venetian election system so complicated?
Each number in the sequence answers a specific problem. In every round except the last, the approvals needed came to the ceiling of three quarters of the college size. That rule has a property the designers would have cared about: no two or three smaller factions could outvote the largest one, and for the college sizes used, not even four could (Mowbray and Gollmann, 2007, pp. 12, 13).
The final threshold of 25 out of 41 breaks the three-quarters pattern. One suggestion is that it was set so that a Doge needed approval from at least one elector outside the twenty-four families counted as the founders of Venice. An earlier analysis found 25 to be the exact number that makes the result as stable as possible under a well-known majority criterion (Mowbray and Gollmann, 2007, p. 13).
Even the opening double draw has a practical explanation. Drawing thirty from the whole Council and then nine from the thirty looks redundant next to drawing nine directly, until you picture the mechanism. Lots were drawn as balls from an urn, and several hundred balls are hard to mix properly while thirty are easy, so the two-stage draw produced a more genuinely random result (Mowbray and Gollmann, 2007, p. 13). The odd number 41 came from an earlier failure: the election of 1229 tied in a college of forty, was settled by drawing lots, and the rules were changed to make the final college odd (Mowbray and Gollmann, 2007, p. 13).
Who could vote, and from what age
The pool was the Great Council, the sovereign assembly of the Republic. Membership was open to all male members of patrician Venetian families over twenty-five, regardless of their individual wealth or standing, and between 1,200 and 2,000 nobles sat in it (Palazzo Ducale, 2026). The Council had grown from 480 members when it was instituted in 1172 to more than two thousand (Molinari, 2018, p. 2).
The election itself used a narrower pool. The first thirty names were drawn only from Council members aged thirty and above, a detail most popular accounts drop (Molinari, 2018, p. 2; Mowbray and Gollmann, 2007, p. 2). So the age of entry to the Council and the age of eligibility for the dogal lottery were five years apart, and the distinction matters for how small the real electorate was.
In 1268 there were 480 Venetian oligarchs (Mowbray and Gollmann, 2007, p. 5). The list of families entitled to sit was closed in 1297, which made it much rarer for new families to enter, and 532 families in total held Council seats between that closure and the end of the Republic (Mowbray and Gollmann, 2007, p. 4). The procedure was elaborate, and the body it drew from was small and largely fixed.
The balotino: the boy who drew the lots
A system built on random draws depends entirely on whoever performs them. Until 1328 an appointed noble drew the lots. That rule was then changed, presumably because the risk of an unfair draw had been noticed, and the task passed to a balotino: a boy chosen as the first boy seen by the youngest member of the inner council of state after that official finished praying at St Mark's Basilica on election day (Mowbray and Gollmann, 2007, p. 13).
The title comes from balota, the Venetian word for the ballot ball used in the election, and the English word ballot descends from the same root (Mowbray and Gollmann, 2007, p. 13). The selection ritual has an obvious flaw, which the analysis identifies: because the place and time were known to everybody in advance, a faction could train a boy to favour it and arrange for him to be standing there at the right moment.
There is one documented case of the voting itself being played. In 1423 an underdog candidate took 17 approval votes of 41 in the ninth ballot of the final college and 26 in the tenth, allegedly because his backers had supported a candidate nobody wanted and then switched. At that date names were still put to the college one at a time, which is what made the manoeuvre possible; concurrent voting on all names came in later (Mowbray and Gollmann, 2007, p. 4).
Did it work? What 75 elections show
Seventy-five Doges were elected under this protocol, and between them they carry only 44 surnames. Seventy of those 75 came from families that already held a Great Council seat in 1297, and over a third came from just thirteen of the twenty-four founding families (Mowbray and Gollmann, 2007, p. 4). The procedure blocked any one faction from arranging an outcome. It did nothing to widen the group the Doge came from.
Where the change is measurable is in dynasty and in tenure. Before 1268, seven of 43 Doges came from the previous Doge's family, and in six of those cases the office passed from father to son. After 1268 it happened twice in 75 elections, and in both cases the two Doges were brothers (Mowbray and Gollmann, 2007, p. 12).
The office was held for life, yet lives in office were short. The 75 Doges elected under the protocol served a mean of 6.85 years and a median of 6, against 11.35 and 11 for the 43 Doges before 1268, and against 19.85 and 14 for the English monarchs whose reigns began in the same period (Mowbray and Gollmann, 2007, p. 11). Electors appear to have preferred older candidates, which limited how long a life appointment could last. The job was also a burden: the Doge paid state expenses from his own pocket, under laws that stopped him profiting from the role (Mowbray and Gollmann, 2007, p. 4).
Where the elections happened in the Doge's Palace
Two connected rooms on the first floor carry the whole story, and both are on the standard museum route. The Chamber of the Great Council is 53 metres long and 25 metres wide, one of the largest rooms in Europe, and it held the assembly the opening lottery drew from (Palazzo Ducale, 2026).
Its decoration dates from the rebuilding after the fire of 1577. Behind where the Doge's throne stood is the Paradiso, painted by Jacopo Tintoretto and his workshop between 1588 and 1592 and described by the museum as one of the longest canvas paintings in the world. Running below the ceiling is a frieze with portraits of the first 76 Doges (Palazzo Ducale, 2026). A visitor standing in the room can therefore see, in one glance, most of the men the procedure produced.
The portraits of the later Doges hang next door in the Chamber of the Scrutinio, and that room is the more exact address for the election. Built in the fifteenth century to house manuscripts left to the Republic by Petrarch and Bessarione in 1468, it was used for electoral procedures from 1532, and in time that became its only purpose. It too was restored after the 1577 fire, between 1578 and 1615 (Palazzo Ducale, 2026).
