Penalties in TM sequences
When you add TMs to a TM sequence, you can define penalties that will be applied to the matches from those TMs.
You can configure penalties for each TM within the TM sequence. As an example, if you have a TM for a related, but different domain compared to the one you are translating into, you can apply penalties to prevent 100% matches from coming from that TM. Alternatively, if you have a TM that contains translations for a slightly different dialect than the current one, you can also apply penalties to prevent 100% from coming from that TM.

TM penalties are leverage-level penalties; they are applied to matches as a whole, as opposed to being applied to numbers, words, or placeholders within segments. If this penalty is applied, you cannot repair the match up to 100%.
TM penalties are leverage-level penalties; they are applied to matches as a whole, as opposed to being applied to numbers, words, or placeholders within segments. If this penalty is applied, you cannot repair the match up to 100%.
You can assign penalty values from 0 to 100. If you specify 2.0, for example, the final score is reduced by two percentage points. In this example, the maximum score that a match from this TM could generate is 98%.
Such penalties are applied only in the context of the TM sequence. If a TM is used in a separate TM sequence, the value of the penalty does not carry over from another TM sequence. Moreover, if you leverage the TM directly, instead of through a TM sequence, the penalty is not applied.