Water is a byproduct of combustion it is the chemical reaction of oxygen with the hydrogen freed from the hydrocarbon chains during combustion. How could the injection of additional present water before combustion contribute further to the combustion process? This is where it gets a bit complex and I will try to do the best I can. This pieces together various pieces of my understanding from different sources.
During early combustion when the fastest reactions occur the effect of the water in the mixture is to cause a more controlled and stable flame front. The freeing of hydrogen and carbon to combine with oxygen has to work around the present water to form OH radicals and CO. By slowing this early combustion process there is further suppression of the potential for the mixture to burn too fast and contribute to knock.
Later in the combustion process slower and more complex reactions occur. The formation of OH radicals is very fast and interferes with the completion of the combustion process to form CO2 from the first step creation of CO. It is during this phase of combustion where present water helps to complete the slower reaction to complete the formation of CO2 since water is about the only way to complete the oxidation of CO. The additional present water actually speeds this reaction which also happens to be when as much of two thirds of the energy from carbon combustion is released.
As additional evidence of why using fuel to suppress knock is a horribly inefficient exercise consider the following:
"...thus one can conclude - correctly - that hydrocarbons inhibit the oxidation of CO.
It is apparent that in any hydrocarbon oxidation process CO is the primary product and forms in substantial amounts. However, substantial experimental evidence indicates the oxidation of CO to CO2 comes late in the reaction scheme. The conversion to CO2 is retarded until all the original fuel and intermediate hydrocarbon fragments have been consumed."
* Source: Combustion, Third Edition, Glassman, p. 76
Fuel dumping necessitates overly advanced timing resulting in timing induced knock (and as has been experienced by others engine destruction) because the significant power release from combustion is delayed to the point where MBT is very late while waiting for the excess fuel is broken down during the extended initial stages of combustion. In upgraded turbo applications this becomes extremely apparent as the increased air flow necessitates ever higher fuel flows to maintain these supposedly safe AFRs of 11:1 and richer.
As illustrated by Sir Ricardo below water injection can actually help reduce maximum cylinder pressures while increasing BMEP. This is because a properly mixed AFR of 12.5:1 accompanied by water injection increases the rate of the combustion reaction exactly when it is desirable to do so during the oxidation of CO => CO2.
Sir Ricardo's results:
*Source: The High-Speed Internal Combustion Engine by Sir Harry Ricardo